audit: track 153 commits lost in bulk revert c3ff635 for recovery

[Audacity88]

Summary

Commit c3ff635 (2026-03-28) reverted 153 commits in a single bulk rollback, restoring master to 1362d69. This revert was necessary at the time, but it also removed bug fixes, features, and improvements that had been reviewed, approved, and merged.

This issue tracks which commits have already been independently re-applied during the microkernel refactor and which are still missing and worth recovering.

Known recovery issues: #5259 (WeChat iLink channel), #2503 (NapCat/OneBot channel)

Note on deleted PRs: Some PR numbers referenced in commit messages return 404 on GitHub. A co-author on several of these commits (SimianAstronaut7, who also authored the revert) had their account blocked, which removed associated PRs and issues. Some external contributor PRs that list SimianAstronaut7 as co-author are also missing — likely opened from that account's fork. Where the PR record is missing, the entry links to the commit instead.

Status key

• ✅ Re-applied — verified present in the current codebase
• 🔲 Missing — verified absent, candidate for recovery
• 🔀 Superseded — the underlying code was refactored or the bug no longer applies in the new architecture
• ⏭️ Skip — merge commit, dep bump, or whitespace-only

---

Completed recovery / coverage PRs (5)

• fix(runtime): persist model_switch across turn paths #6719 — 58353504 feat(channels): persist per-chat model switch to routes.json #4648 model_switch recovery merged as e147e491. ✅
• feat(channel/mattermost): add WebSocket listener mode #9141 — 78833057 feat(mattermost): add WebSocket-based real-time listener #4845 Mattermost WebSocket listener recovery merged as 0de7f340. ✅
• feat(runtime): add calendar no-show SOP triggers #8419 — 3f229c85 feat(calendar): add calendar-driven no-show detection triggers #4527 calendar no-show SOP trigger recovery merged as 3278d76f. ✅
• fix(channels): fire message_sent hooks after delivery #8355 — 2dba7fe0 feat(hooks): fire message_sent hook after successful channel delivery #4493 message_sent hook recovery merged as 8c080dd3. ✅
• refactor(ci): consolidate prebuilt Docker image variants #8485 — 359c8006 Docker CI file consolidation recovery merged as 8ac9565f. ✅

Verified re-applied (77)

• 4eb9cbe4 feat(nodes): add mDNS local peer discovery #4428 — mDNS/local peer discovery (re-applied by feat(gateway): add LAN peer discovery hints #8325 / 344a158: current opt-in [nodes.mdns] publishes runtime-derived gateway hints, bounds and expires unauthenticated LAN peers, exposes them through authenticated status without auto-connecting or bypassing node registration.) ✅

These fixes have been independently re-implemented in the current crate structure:

• 4022879b feat(channels): add room creation and user invite API to Channel trait #4504 — Matrix room creation + invite recovery (re-applied by feat(channels): restore Matrix room management tool #8068 / Matrix room-management tool recovery: typed Matrix room-management APIs plus the agent-facing channel_room path are now merged. No live Matrix homeserver smoke test was claimed in the recovery row.) ✅
• 99f0b528 feat(channels): add /thinking runtime command for per-sender reasoning toggle #4325 — per-sender /thinking channel override recovery (re-applied by feat(channels): restore per-sender /thinking overrides #8011: current channel command support restores the user-facing override on top of ThinkingLevel plumbing instead of the old provider-clone reasoning_enabled path.) ✅
• ac680041 feat(security): add shared SsrfValidator with CIDR blocking and homograph detection #4421 — http_request SSRF DNS-answer validation / request pinning (re-applied by fix(tools): pin http_request requests to vetted DNS addresses #7902: current http_request pins requests to vetted DNS addresses while leaving the old unwired SsrfValidator API and custom-CIDR config surface behind.) ✅
• 7f08d4b0 feat(service): detect missing loginctl linger and prompt user #4285 — detect missing loginctl linger + prompt (re-applied/evolved by fix(runtime): warn when systemd user lingering is disabled #8249 and the current installer/systemd preflight path: missing linger state is detected before service setup instead of surfacing as a later runtime/service failure.) ✅
• df129fea fix(service): detect missing loginctl enable-linger and warn users #4336 — detect missing loginctl enable-linger (re-applied/evolved by fix(runtime): warn when systemd user lingering is disabled #8249 and the current installer/systemd preflight path, which reports the missing enable-linger requirement explicitly.) ✅
• 1ff39882 fix(service): detect missing loginctl enable-linger on install #4435 — detect enable-linger on install (re-applied/evolved by fix(runtime): warn when systemd user lingering is disabled #8249 and the current installer/systemd setup checks.) ✅
• 20238a26 (commit) — Rust version check for build failures (covered by the current install.sh / installer rewrite path, which validates the Rust toolchain before relying on source builds.) ✅
• 961d13ad (commit) — rate-limit cooldown + model compatibility (re-applied by fix(provider): cool down rate-limited fallback entries #8317 / 6d2ce03: current provider fallback entries cool down after rate-limit responses instead of immediately retrying incompatible or exhausted routes.) ✅
• 56c7d605 feat: add Tavily as configurable web search provider #4685 — Tavily web search provider (covered by feat(tools/web_search): implement Tavily search backend #6185 and the current web search provider/tool routing; the old standalone row is no longer a missing recovery target.) ✅
• dd147dd0 feat: add zeroclaw config reload for hot-reloading config #3571 — zeroclaw config reload / hot-reload (superseded by the current admin/RPC reload path and live config reload wiring, not the old monolithic command shape.) ✅
• f685de05 fix(tests): serialize Bedrock env-var tests to prevent parallel race #4883 — serialize Bedrock env-var tests (covered by current env_lock() / EnvGuard test utility in zeroclaw-providers; Bedrock env-var tests now take the shared lock before mutating process environment.) ✅
• 74a211a5 test(pushover): add unit tests and improve documentation #3830 — pushover unit tests + docs (covered by the current zeroclaw-tools Pushover unit tests plus test(tools): cover pushover notification shape overlap #8356, including schema overlap, credential parsing, priority/sound, read-only, and rate-limit behavior.) ✅
• 20c589ee ci: add deployment environments to release workflows #4839 — deployment environments in release workflows (re-applied/evolved by ci: add deployment environments to release workflows #4839/feat(ci): add GitHub deployment environments #5073 and current release-stable-manual.yml environment-gated publish/redeploy jobs.) ✅
• 2d73133d (commit) — feishu transport media (re-applied by feat(channels/lark): restore outbound media markers #8113 / b980e3a: current Lark/Feishu outbound sends parse shared [IMAGE:], [PHOTO:], [DOCUMENT:], [FILE:], [VIDEO:], [AUDIO:], and [VOICE:] markers and route marker-derived attachments through the current Lark upload/send path.) ✅
• 58ad2092 feat(knowledge): add client relationship node types and tool actions #4507 — client relationship node types (re-applied by feat(knowledge): restore client relationship graph actions #8182 / c6906da: current knowledge graph restores Client, Contact, and Interaction node types, client relationship relations, graph_neighbors, client_network, and interaction_log knowledge tool actions, Postgres conversion, and localized tool descriptions.) ✅
• 2ddf7dcb fix(agent): add web_search_tool to default_param_for_tool mapping #4544 — web_search_tool in default_param_for_tool mapping (tool-call-parser/lib.rs:807) ✅
• 49e994a3 fix(channels): handle list message type in Feishu/Lark channel #4702 — Lark list message type handling (lark.rs:986, test at :2766) ✅
• ffb015b3 fix(security): make command allowlist matching case-insensitive on Unix #4552 — case-insensitive command allowlist matching (policy.rs:824) ✅
• 143b34ea (commit) — SearXNG search provider config (schema.rs:2977) ✅
• a4e6d772 feat(slack): migrate audio transcription to TranscriptionManager #4314 — Slack audio file transcription (slack.rs:39) ✅
• a3da42b1 — mention_only arg + voice_wake feature-gated errors (telegram.rs:367) ✅
• 2a5866ec fix(update): extract binary from .tar.gz release assets #4297 — extract binary from .tar.gz release assets (update.rs:212) ✅
• 490c54e8 fix(security) allow high risk command execution #4750 — allow high risk command execution (whitespace-only, no functional change) ✅
• 8416af7d (commit) — local time for runtime_trace timestamps (production path at line 216 uses Local::now(); remaining Utc calls are temp file naming and tests only) ✅
• 1e477f52 (commit) — respect max_history_messages in channel mode (orchestrator/mod.rs:1211) ✅
• a993c96f fix(agent): use stream_chat in turn_streamed to fix WS/webhook streaming #4690 — use stream_chat in turn_streamed for WS/webhook streaming (loop_.rs:547, ws.rs:467) ✅
• 3b9c184e fix(provider): use no_responses_fallback for Z.AI and add Bailian alias #4317 — Z.AI no_responses_fallback + Bailian alias (provider_aliases.rs:89, lib.rs:1293) ✅
• e9c99f92 fix(matrix): preserve thread context on first follow-up message #4805 — preserve matrix thread context on first follow-up (matrix.rs:256-292) ✅
• b7f3d0c9 feat(matrix): add mention_only config for group room filtering #4680 — Matrix mention_only for group room filtering (re-applied by feat(matrix): clean-room rewrite on matrix-rust-sdk 0.16 #6112 / ce3a968: current MatrixConfig.mention_only gates group rooms through is_group_room plus mention::is_mentioned, using m.mentions.user_ids when present and full MXID/localpart/display-name body fallback otherwise; direct rooms bypass via room.is_direct(). Focused validation passed: cargo test -p zeroclaw-channels mention --features channel-matrix -- --nocapture with 65 passed / 0 failed. The old string-strip helper is superseded by the Matrix SDK rewrite rather than carried as a separate preprocessing step.) ✅
• 959057ab feat(tools): wire session tools to composite backend for gateway visibility #4852 — wire session tools to composite backend for gateway visibility (re-applied by fix(runtime, channels): unify session backend behind one factory #6384 / c6a0ed7: current runtime session tools, channel orchestrator, and gateway /ws/chat persistence all call zeroclaw_infra::make_session_backend, so [channels].session_backend is the single configured store and gateway sessions are visible to sessions_list, sessions_history, and sessions_send. The old read-time CompositeSessionBackend wrapper is superseded by the factory plus SQLite-default JSONL import path. Focused validation passed: cargo test -p zeroclaw-infra make_session_backend --lib -- --nocapture with 4 passed / 0 failed.) ✅
• 96643394 feat(channels): precheck reply intent before typing or draft send #4850 — precheck reply intent before typing/draft (re-applied in current channel orchestration: classify_channel_reply_intent / parse_reply_intent classify visible-reply intent, the no-reply gate runs before draft creation and typing setup in process_channel_message_body, and no-reply outcomes record a history marker plus reaction/log evidence. Focused validation passed: cargo test -p zeroclaw-channels precheck --lib with 2 passed / 0 failed.) ✅
• 3ce90480 (commit) — respect path_prefix in get-paircode command (gateway routes nested under prefix at lib.rs:1100) ✅
• 516a9cfb fix(config): resolve temp directory canonicalization for non-existent… #4529 — resolve temp dir canonicalization on macOS (schema.rs:9418, seatbelt.rs:151) ✅
• a279f5da fix(gateway): improve web dashboard unavailable message for Homebrew users #4438 — improve dashboard unavailable message for Homebrew (service/mod.rs:582-620) ✅
• 9ce1b67a (commit) — prevent duplicate tool call notifications (superseded by fix(agent): drop redundant narration push before AssistantToolCalls #6093; dedup_exempt infra in schema.rs:1487) ✅
• 15a630e4 fix(gateway): clarify public bind error message #4763 — clarify public bind error message (gateway/lib.rs:413-422) ✅
• 6044fa23 fix(tools): resolve agent-browser npm shim on Windows #4557 — resolve agent-browser npm shim on Windows (browser.rs:247-250) ✅
• 59c9a2d1 fix: resolve claude-code test flakiness and update security policy #4181 — resolve claude-code test flakiness (superseded by fix(agent): drop redundant narration push before AssistantToolCalls #6093; agent.rs:2400-2410) ✅
• ac355d40 feat: add GitHub copilot to onboarding #4897 — add GitHub Copilot to onboarding (re-merged as feat: github copilot onboarding #5321, 9f245539f) ✅
• 20593311 (commit) — enable MCP tools in webhook endpoint (re-applied by fix(gateway): enable tool support in webhook endpoint #6080, be106b681) ✅
• 59ad6da0 fix(memory): add closing tag to memory context injection #4687 — add closing tag to memory context injection (re-applied by fix(memory): add closing tag to memory context in cron and daemon #6082, 3b24f817) ✅
• 3e0a08c1 (commit) — add missing bind paths for bubblewrap sandbox (re-applied by fix: recover 4 small fixes lost in bulk revert c3ff635 #6169, 62728c1ceb) ✅
• df43e516 fix(gateway): respect path_prefix in get-paircode and shutdown commands #4555 — respect path_prefix in paircode + shutdown commands (re-applied by fix: recover 4 small fixes lost in bulk revert c3ff635 #6169, 62728c1ceb) ✅
• 8c3c890c (commit) — add timeout to Mattermost HTTP client (re-applied by fix: recover 4 small fixes lost in bulk revert c3ff635 #6169, 62728c1ceb) ✅
• babe7f1e fix(heartbeat): inject current datetime into scheduler decision prompt #4554 — inject current datetime into scheduler decision prompt (re-applied by fix: recover 4 small fixes lost in bulk revert c3ff635 #6169, 62728c1ceb) ✅
• 74a94a1c fix(channels): skip tools summary for native tools #4499 — skip tools summary for native tools (re-applied by fix(runtime): omit native tool prompt catalog #6544, 3821550c) ✅
• 4995c184 fix(update): use exact arch triple matching to prevent wrong-binary downloads #4337 — exact arch triple matching for update downloads (re-applied by fix(update): tighten release asset selection #6585, ed9bd52) ✅
• 4f872cca fix(provider): restrict reasoning_effort to known openai models (#4296) #4310 — restrict reasoning_effort to known openai models (re-applied by fix(provider): restrict compatible reasoning effort models #6631, d2dd45a) ✅
• 7882897e fix(gemini): propagate token usage from Gemini provider to cost tracker #4573 — propagate Gemini token usage to cost tracker (re-applied by fix(gemini): propagate token usage to cost tracker #6575, f8221ce) ✅
• e7467e0a fix(provider): fallback to non-streaming when OpenAI Codex streaming fails #4538 — fallback to non-streaming when Codex streaming fails (re-applied by fix(provider): recover Codex non-streaming fallback #6668, ea825ce) ✅
• d4ab8a47 feat(debug): add --log-llm flag to dump LLM provider message payloads #4213 — add --log-llm flag to dump LLM provider payloads (re-applied by feat(debug): restore --log-llm payload tracing #6709, 3e27b58) ✅
• da742938 fix(multimodal): gracefully skip unresolvable images instead of failing #4316 — gracefully skip unresolvable images instead of failing (re-applied by fix(providers): skip unresolvable multimodal images #6743, f6b90124) ✅
• 46ca9364 fix(onboard): restore --interactive flag for TTY override #4410 — restore --interactive CLI compatibility; current bare onboarding is already interactive (re-applied by fix(onboard): restore interactive flag compatibility #6673, f1ced29e) ✅
• ef9e52f3 fix(security): respect wildcard opt-out for subshell/redirect guards #4803 — respect wildcard opt-out for subshell/redirect guards (re-applied by fix(shell): guard bypass #5773, 247e2e18) ✅
• 5b6d0585 Feat/vision support for kimi code #4108 — enable vision support for kimi-code provider (re-applied by fix(providers): restore Kimi Code vision capability #6809, 7f9271cb) ✅
• ce6dd7ca fix(security): mask Groq API key and transcription provider keys in web UI #4688 — mask Groq API key + transcription keys in web UI (web config masking is already covered by current schema-derived secret masking; remaining Groq leak-detector pattern re-applied by fix(security): detect Groq API keys in leak scanner #6812, 84413ff7) ✅
• aabf2327 fix(ollama): preserve :cloud model tag for private remote Ollama servers #4173 — preserve :cloud model tag for private remote Ollama servers (re-applied by fix(providers): preserve Ollama cloud suffix for private remotes #6863, 9c2522c4) ✅
• ecd68feb fix(onboarding): detect Homebrew install and use correct config path #3526 — detect Homebrew install, use correct config path (re-applied by fix(onboarding): restore Homebrew config dir detection #6896, cc1a66b) ✅
• f12c1761 fix(providers): add streaming support to RouterProvider #4533 — add streaming support to RouterProvider (remaining system-streaming delegation re-applied by fix(providers): route system streaming through RouterModelProvider #6935, f4d17ea; current master already covered supports_streaming, stream_chat_with_history, and structured stream_chat) ✅
• c96bd63e fix(update): verify download checksum against SHA256SUMS #4434 — verify download checksum against SHA256SUMS (re-applied by fix(update): verify self-update downloads against SHA256SUMS #6900, ec4d0f7) ✅
• a68f2dd3 fix(channels): close draft sender after tool loop #4301 — close draft sender after tool loop (current orchestrator drops delta_tx before awaiting the draft updater at crates/zeroclaw-channels/src/orchestrator/mod.rs:3422; original Telegram draft-hang regression is not present) ✅
• 6469a262 fix(providers): unify OpenAI-compatible native tool response parsing #4149 — accept compatible typed content arrays (re-applied by fix(provider): accept compatible typed content arrays #7010, 58c24e2c) ✅
• 31737b16 fix(prompt): timestamp channel turns and use date-only prompts #3271 — timestamp channel turns, use date-only prompts (re-applied by fix(channels): restore date-only channel prompt context #6931, 9651e13) ✅
• 371c59d1 (commit) — unify path resolution for file tools (remaining image_info path-resolution side re-applied by fix(tools): resolve image_info paths through policy #6972, 708f1f6) ✅
• b58d18d5 fix(agent): fallback to non-streaming when stream errors (#4670) #4675 — fallback to non-streaming when stream errors (re-applied by fix(runtime): fall back before visible stream errors #6983, 4ed0334) ✅
• 51fd8ba5 feat(qq): add TranscriptionManager wiring for audio attachments #4315 — QQ audio attachment transcription (re-applied by fix(qq): restore audio attachment transcription #7109, f8641d2) ✅
• ad48c1de feat(config): add configurable global max_audio_bytes to TranscriptionConfig #4114 — configurable global max_audio_bytes (re-applied by feat(config): restore transcription max_audio_bytes #7303, 81ce4f91) ✅
• 753d4fc6 feat(tool): add SecretStore integration to http_request #3637 — SecretStore integration in http_request tool (re-applied by feat(tools): add http_request auth secrets #7354, d064ac36) ✅
• 21e23bd2 feat(channel): add interrupt_on_new_message support for WhatsApp #4371 — interrupt_on_new_message for WhatsApp (re-applied by fix(channel): restore WhatsApp interrupt_on_new_message #7040, 67acda2) ✅
• a54e9257 feat(security): add 1Password secret resolution via op:// references #4796 — 1Password secret resolution via op:// references (re-applied by fix(config): restore 1Password op secret refs #7085, f3c8c77) ✅
• d07f4208 feat: support forwarded messages in Telegram channel (#4118) #4254 — Telegram forwarded-message attribution (re-applied by fix(channels/telegram): restore forward origin attribution #7251, feb5776) ✅
• 49620636 feat(voice): add unified VoicePipeline facade for STT+TTS channels #4427 — voice pipeline facade for shared channel STT/TTS use (re-applied by feat(channels): add voice pipeline facade #7394, 4169c163) ✅
• 4fe1f1fd feat(matrix): populate media attachments and add outgoing media support #4865 — Matrix media attachments + outgoing media (re-applied/superseded by feat(matrix): clean-room rewrite on matrix-rust-sdk 0.16 #6112 / ce3a968: current Matrix persists inbound media into workspace-backed media markers via attach_media / with_workspace_dir, parses outbound [IMAGE:] / [DOCUMENT:] / [VIDEO:] / [AUDIO:] / [VOICE:] markers through the shared channel marker parser, uploads marker-derived bytes and structured SendMessage.attachments with thread anchoring, and feat(matrix): upload attachments when finalizing partial drafts #6200 / b699f5d plus fix(matrix): include attachment media metadata #6610 / 3afb545 hardened partial-draft attachment finalization and attachment metadata. Source/provenance check only in this pass; focused local cargo was not run because the SSD Cargo target root was unavailable and internal-disk fallback was not approved.) ✅
• b69e5ff5 fix(tools): improve file_read robustness for large/binary files #4320 — reject binary/image files in default file_read instead of returning lossy text (re-applied by fix(tools): reject binary/image files in default file_read instead of lossy text #7522 / 91b0bd7: current default UTF-8 reads reject image signatures and confident binary content with guidance to use image_info or encoding="base64", while explicit base64 reads and conservative non-UTF-8 text handling remain available.) ✅
• 798aa936 (commit) — native Firejail/Bubblewrap hardening (re-applied/superseded by fix(runtime): restore native sandbox hardening #7683 / 34107bb3: current native sandbox wrapping adds supported Firejail seccomp/capability/noroot hardening and Bubblewrap capability drops behind backend support detection, while deliberately avoiding the old invalid bare Bubblewrap --seccomp path and not restoring the older unavailable-backend fail-closed attempt.) ✅
• 64043d64 feat(channels): parse media attachment markers in WhatsApp Web send #4403 — parse media markers in WhatsApp Web send (re-applied by feat(channels/whatsapp): send Web media markers natively #7811 / b1831f9: current WhatsApp Web outbound sends parse shared [IMAGE:], [PHOTO:], [DOCUMENT:], [FILE:], [VIDEO:], [AUDIO:], and [VOICE:] markers, constrain local marker paths to the configured WhatsApp workspace, and send native WhatsApp image/video/audio/voice/document messages through the current whatsapp-rust upload/send API.) ✅

Checked / superseded small fixes (4)

• 44eb057f fix(gateway): replay session history on WebSocket reconnect #4696 — replay session history on WebSocket reconnect (superseded by feat(web): persist Agent Chat history across navigation and refresh #4792 / 5c7eccb and feat(webui): hot-switch model & preserve chat context across page navigation #6101 / d145a24: current master exposes GET /api/sessions/{id}/messages, and AgentContext hydrates chat from that endpoint before the shared WebSocket takes over) 🔀
• 39891105 fix: issue 4855 #4895 — fix [Bug]: Channels API not working #4855 / Channels API returning HTML instead of JSON (covered by fix(gateway): add missing /api/channels route #6069 / eebd7b6: current master routes GET /api/channels; the dashboard helper also accepts the current wrapped { "channels": [...] } response shape) ✅
• 933f7a25 fix: sessions page #4858 — fix sessions (re-applied by bf7380b5) ✅
• 9fda248c (commit) — preserve web chat history across page reloads (superseded by the current GET /api/sessions/{id}/messages API plus AgentContext hydration/localStorage fallback; the backend also sends session_start and seeds the agent with stored messages on WebSocket resume) 🔀

Checked / superseded medium fixes (13)

• a51620b4 fix(prompt): expose autonomy constraints to the model #3296 — expose autonomy constraints to the model (superseded by the current risk-profile prompt path: fix(prompt): respect autonomy level in SafetySection (Agent/gateway WS path) #4037 / 96f25ac7 wires autonomy level into SafetySection, fix(prompt): rewrite safety section to reduce AI conservatism #4615 / 671577b4 keeps the model-facing wording direct, and feat!: multi-agent runtime and schema V3 #6398 / 18622d91 routes channel/webhook runs through per-agent risk_profile so prompts include runtime autonomy plus active security-policy constraints instead of the old standalone ## Autonomy Constraints block) 🔀
• 6335a0d4 fix(channel/discord): download image and video attachments for agent processing #4871 — download Discord image/video attachments for agent processing (re-applied by fix(discord): preserve inbound image attachments for providers #6184 / 213b01dd and broadened by fix(channels): close Discord media send/receive gaps #6572 / ccb6beb29: current Discord preprocessing admits attachment-only messages, downloads non-text attachments once, emits IMAGE/VIDEO markers, stores MediaAttachment bytes on ChannelMessage, and the media pipeline classifies image/video MIME types for provider handling) ✅
• 03d345b5 fix(providers): skip responses fallback on transport errors #4501 — skip responses fallback on transport errors (superseded by feat!: multi-agent runtime and schema V3 #6398 / 18622d91b: compatible.rs no longer has chat_via_responses or supports_responses_fallback; chat transport errors return directly, and /responses is isolated to OpenAiCodexModelProvider) 🔀
• 80a91856 fix: prevent panic on UTF-8 multi-byte char boundary in slug truncation #4148 — prevent panic on UTF-8 multi-byte boundary in slug truncation (behavior present in crates/zeroclaw-runtime/src/skills/creator.rs; original CJK regression test not present) ✅
• e2022f9e fix(agent): consolidate multiple messages into single response #4436 — consolidate multiple messages into single response (superseded by current AgentContext/AgentChat tool-activity toggle; tool activity is hidden from the conversation by default) 🔀
• f17c6dce (commit) — disable responses API fallback for custom OpenAI providers (superseded by the current provider split: custom: uses chat completions through OpenAiCompatibleProvider; /v1/responses is handled by OpenAiCodexProvider / wire_api = "responses" profiles, not a 404 fallback path) 🔀
• c74824a3 fix(tools): make shell timeout configurable via config.toml #4334 — make shell timeout configurable via config.toml (covered by later feat: make shell tool timeout configurable via config.toml #4468 / ea9eccfe8; current master exposes [shell_tool].timeout_secs, wires it through the tool registry, and applies it in ShellTool execution) ✅
• 735a1336 feat(ollama): allow configurable context size via ZEROCLAW_OLLAMA_NUM_CTX #3518 — configurable Ollama context size (superseded by feat(providers/ollama): add num_ctx/num_predict/temperature tuning #6178 / e6cbb33d; current master exposes config-backed ollama_num_ctx, ollama_num_predict, and temperature override) 🔀
• f78a6d32 fix(agent): separate draft tool narration lines #4349 — separate draft tool narration lines (covered by fix(channels): ensure newline between narration and draft status lines #4394 / 0d2b57ee; current master still appends a missing newline before native tool-call narration is sent to draft-capable channels, with runtime regression coverage) 🔀
• 6aab1288 fix(channels): add channel-lark to default features to support Lark/Feishu long connection #4444 — add channel-lark to default features (covered by current workspace feature graph: root default enables agent-runtime, agent-runtime enables channel-lark, and channel-lark forwards to zeroclaw-channels/channel-lark; Docker builds do not use --no-default-features, so the same default applies there) 🔀
• 41c0abbf (commit) — whitelist safe shell redirects (covered by fix(security): allow safe device redirect targets in path policy #6236 / ad34fa3f; current policy allows safe /dev redirects and fd redirects while still blocking file redirects and /dev path-boundary bypasses, with focused redirect-policy tests) ✅
• 88d77ce3 fix(channels): prevent draft streaming hang after tool loop completion #4359 — prevent draft streaming hang after tool loop completion (covered by fix(channels): prevent draft streaming hang after tool loop completion #4393 / b85a44595; fix(channels): prevent draft streaming hang after tool loop completion #4359’s commit carries no file diff, and current orchestrator drops delta_tx before awaiting draft_updater so the draft update task can observe channel close and exit) 🔀
• 3c592593 (commit) — bound conversation history with LRU sender cap (re-applied by 709b867f) ✅

Checked / superseded large fixes (4)

• 9528678e (commit) — trim oldest images instead of hard-failing when limit exceeded (re-applied by 541b273b) ✅
• 61c32424 (commit) — preserve existing channel config when re-configuring (re-applied by 351986c2) ✅
• dc05efa9 fix(channels): use reversible percent-encoding for session file keys #4811 — reversible percent-encoding for session file keys (superseded by fix: sanitize session keys at the channel orchestrator layer #6228 / 388657f68: current channel orchestration pre-sanitizes conversation history and memory keys with shared sanitize_session_key before they reach the JSONL/session backends, so hydration, list, load, memory recall, and runtime cache lookup all use the same normalized key. Current SQLite session metadata also stores channel_id, room_id, and sender_id separately, so original Matrix/platform IDs no longer depend on reversible filenames; fix(channels): scope session key for channel tools #6541 / 247f2baf5 keeps channel session tools scoped to that same history key.) 🔀
• 87698ad1 fix(agent): treat tool_use/tool_result as atomic groups in history pruning #4825 — treat tool_use/tool_result as atomic groups in history pruning (re-applied by fix(session): session integrity, streaming refactor, history pruning #5167 / 1ec9c14c and hardened by fix(agent): drop orphan tool_results when trimming history #5485 / d72c2128, fix(runtime): self-heal orphaned tool_result blocks on load + compact #5853 / a449a506c, and fix(agent): treat mixed-protection tool groups as atomic in prune_history Phase 1 #5828 / 3d7f2ee0: current prune_history collapses and drops assistant+tool groups atomically, emergency_history_trim does the same, trim/load/channel hydration self-heal orphan tool results, and regression coverage includes multi-tool collapse/drop, realistic token pressure, protected recent groups, mixed-protection groups, and structured orphan cleanup.) ✅

Checked / superseded features (22)

• c373d6c9 feat: add movie extension with Douban and TMDB support #4815 — movie extension with Douban + TMDB (superseded as a recovery row: the old extension tree is gone, and replaying the monolithic movie extension would not fit the current crate/plugin shape. A movie/media lookup feature should be scoped fresh if the project still wants it.) 🔀

• 070a0c20 feat: adopt AGENTS.md as primary agent instruction format #4290 — adopt AGENTS.md as primary instruction format (partially recovered/evolved by the current root/project AGENTS.md guidance and local wrapper-skill model. The old crate-layout assumptions and gates should not be blindly replayed.) 🔀

• f657bba4 feat(channel): add WeChat iLink Bot channel #4221 — WeChat iLink channel (tracked separately by [Feature]: recover developed WeChat iLink channel in reverted commits #5259, so this tracker should not duplicate the WeChat recovery row here.) 🔀

• e5e4d27f feat(agent): add native_tool_calls_only config to disable text fallback parsing #3682 — native_tool_calls_only config (superseded by fix(runtime): add strict tool parsing mode #6675 / 7437cd0e: current master uses per-agent/runtime-profile strict_tool_parsing instead of restoring the old global [agent].native_tool_calls_only key. Strict parsing suppresses text fallback tool-call execution for direct agent turns, streamed/channel turns, and delegate subagents according to the active target-agent policy; the no-text-fallback local-agent profile is documented. Source/provenance check only in this pass; fix(runtime): add strict tool parsing mode #6675 carried focused and broad validation before merge.) 🔀

• 2252c3bb feat(security): add per-channel DM pairing manager #4530 — per-channel DM pairing manager (superseded by the current channel-specific pairing architecture: PairingGuard now lives in zeroclaw-config with hashed token persistence, brute-force lockout, code-regeneration controls, and focused tests; Telegram, WeChat, and LINE wire /bind <code> pairing into inbound paths and persist successful identities into channel peer groups via orchestrator-provided config handles. The old standalone DmPairingManager JSON module was foundation-only with no downstream callers, so it should not be restored as a separate security module.) 🔀

• d00d5c28 feat(channel): add observe_group flag and per-chat session keys #4502 — observe_group flag + per-chat session keys (partially superseded by current channel/session architecture: ChannelMessage now derives Default, session keys are sanitized and channel-alias-aware, and Lark has an explicit per_user_session knob that defaults group chats to shared chat sessions. The old cross-channel per-chat key change is not present and should not be restored as a blind recovery because current non-Lark channels still isolate history by sender. The passive observe_group side is also absent: current mention_only paths drop unmentioned group messages rather than storing them, so any revival should be a fresh design decision.) 🔀

• be5e3e9d feat(channels): add message chunker with per-platform character limits #4422 — message chunker with per-platform char limits (superseded by current channel-specific send-path chunking: Telegram uses TELEGRAM_MAX_MESSAGE_LENGTH plus send_text_chunks, Discord enforces DISCORD_MAX_MESSAGE_LENGTH through chunks_for_send, IRC splits PRIVMSG payloads by safe UTF-8 byte limit, Slack splits Block Kit text into bounded sections with truncation, and Lark splits interactive-card markdown by byte limit. The old shared src/channels/chunker.rs helper was unused foundation-only code and was not restored as a standalone module. Focused validation passed from detached origin/master: cargo test -p zeroclaw-channels split --features channel-telegram,channel-lark --lib -- --nocapture with 20 passed / 0 failed, cargo test -p zeroclaw-channels discord::tests::split --features channel-discord --lib -- --nocapture with 28 passed / 0 failed, and cargo test -p zeroclaw-channels irc::tests::split --features channel-irc --lib -- --nocapture with 11 passed / 0 failed.) 🔀

• b67be417 feat(agent): add thinking/reasoning level control per message #4332 — thinking/reasoning level control per message (the commit itself is tree-identical to its parent because fix(agent): prevent thinking level prefix from leaking across turns #4277 / ffb8b81f already carried the feature plus the interactive prompt-restore fix; current master retains the feature in crates/zeroclaw-runtime/src/agent/thinking.rs and crates/zeroclaw-config/src/scattered_types.rs, wires /think:<level> through single-shot, interactive, and channel paths, and feat(provider): add native extended thinking for Anthropic and Bedrock providers #5652 / fe570ce4 broadens the same config with opt-in native extended-thinking budgets.) ✅

• da0b083e feat(config): add provider_env for injecting API keys from config #4322 — provider_env for injecting API keys from config (superseded by typed encrypted provider API-key fields plus schema-derived env overrides; current config stores secondary provider credentials directly under providers.models.<family>.<alias>.api-key, encrypts them via SecretStore, and tests provider/custom-slot secret routing, so the old process-wide [provider_env] table is no longer needed.) 🔀

• 6ab4d03b feat(cron): add Feishu/Lark channel support for cron job delivery #4378 — Feishu/Lark cron delivery channel (superseded by feat(channels,runtime): support lark as a cron delivery channel #6851 / 3498f50e and later channel/runtime structure: current master routes lark and feishu cron announce delivery through crates/zeroclaw-channels/src/orchestrator/mod.rs, exposes both names in cron_add and cron_update, handles the shared [channels.lark.<alias>] config with use_feishu, and includes send-boundary plus route-to-lark-arm regressions.) 🔀

• 58353504 feat(channels): persist per-chat model switch to routes.json #4648 — persist per-chat model switch to routes.json (superseded by the current runtime/session model_switch contract and tracked by [Bug]: model_switch tool does not persist across turns; gateway/UI path ignores it entirely #6173 / fix(runtime): persist model_switch across turn paths #6719. The old disk-backed routes.json daemon-restart persistence is not the current model-switch contract; [Bug]: model_switch tool does not persist across turns; gateway/UI path ignores it entirely #6173/fix(runtime): persist model_switch across turn paths #6719 cover the remaining cross-turn and gateway/UI behavior.) 🔀

• aa419d57 feat(security): add path-validation fallback sandbox #4531 — path-validation fallback sandbox (superseded by current application-layer security: shell/file-capable tools are registered through PathGuardedTool, and SecurityPolicy enforces workspace-only plus comprehensive forbidden_paths even when OS sandbox auto-detection falls back to NoopSandbox. The old standalone PathValidationSandbox backend was not restored as a separate sandbox mechanism.) 🔀

• 002fe4c3 (commit) — upgrade claude-code provider to full agent mode (re-applied by 4919216b) ✅

• ffea087c (commit) — ToolCallCard component for tool call display (re-applied by b08cf79a) ✅

• 5d8d9aa2 feat(channel): add LINE Messaging API channel with Reply/Push API support #4822 — LINE Messaging API channel (re-applied by 6ce9a737) ✅

• 936a7212 (commit) — generic transport stall watchdog (re-applied by 5a5d28f) ✅

• 3e02e68e (commit) — ZhipuJwt auth for Z.AI and GLM providers (re-applied by 34ef7bfe) ✅

• 6a8ff4c4 (commit) — email attachment download + multipart MIME (re-applied by dc7a7be7) ✅

• 3f6e7c5c (commit) — WhatsApp Web mention_only for groups (re-applied by fab350eb) ✅

• f3d8aa1f (commit) — MQTT channel configuration schema (re-applied by c6a980d0) ✅

• bde81df1 (commit) — memory namespace isolation for delegate agents (re-applied by 9537885f) ✅

• 9a3453e6 (commit) — per-sender rate limiting (re-applied by 90edff25) ✅

Checked / superseded tests, infra, CI (10)

• a57145cc fix(homebrew): publish a release source archive with built dashboard assets #3934 — Homebrew release source archive with dashboard (superseded by the current release/install source-dashboard flow. Current package-publishing access failures are release-infra work, not this lost-row recovery target.) 🔀
• 7a5d6837 (commit) — chat history persistence (superseded by current gateway session APIs, WebSocket session resume, and AgentContext/chat workspace hydration instead of the old one-off web chat storage patch.) 🔀
• d4773099 (commit) — web chat history with optimized scroll (superseded by current ChatWorkspace/chatHistoryStorage and session-message hydration behavior.) 🔀
• 248e13eb fix(cron): add WhatsApp Web delivery channel for cron job announcements #4258 — WhatsApp Web cron delivery channel (covered/evolved by current cron delivery routing and fix(channels): support WhatsApp as a cron delivery channel #7562, which supports WhatsApp as a cron delivery channel on the current runtime/channel path.) 🔀
• 65ae6cb7 fix(cron): add live channel registry for WhatsApp Web cron delivery #4548 — live channel registry for WhatsApp Web cron (covered/evolved by the current cron delivery enum/registry tests plus fix(channels): support WhatsApp as a cron delivery channel #7562; no separate old registry patch should be replayed.) 🔀
• bedbbdcb (commit) — unify env-var test sync across providers (re-applied by 92a98959) ✅
• fbc654da (commit) — extract rate-limiting + path-guard wrappers (re-applied by 85f9e6a8) ✅
• f0ffb2b5 (commit) — WhatsApp config regression test without [cli] (covered by current tests/component/config_schema.rs: config_only_whatsapp_channel_parses exercises [channels.whatsapp.default] with no [cli] and asserts channels.cli defaults true. Source/provenance check only; no local cargo run in this pass.) ✅
• 1a1a27e1 fix(config): add regression test for WhatsApp config without [cli] field #4432 — WhatsApp config regression test without [cli] (duplicate/no-diff follow-up for the same [Bug]: Permissions issue trying to start WhatsApp #3456 regression; covered by current tests/component/config_schema.rs WhatsApp-without-[cli] regression coverage. Source/provenance check only; no local cargo run in this pass.) ✅
• d781fb5b fix(tests): wrap LocalWhisperConfig.bearer_token test values in Some(...) #4734 — wrap LocalWhisperConfig.bearer_token fixtures in Some(...) (commit is tree-identical/no-diff in the recovered chain; current LocalWhisperConfig.bearer_token is Option<String> and current LocalWhisperConfig fixtures use Some(...) / None consistently. Source/provenance check only; no local cargo run in this pass.) ✅

Skipped (17)

Merge commits, dependency bumps, docs-only, chore commits, and items whose target files no longer exist — not worth recovering individually.

• a50e2304 (commit) — update blob/main→blob/master link in movie extension docs (file extensions/movie/CONTRIBUTING_GUIDE.md no longer exists in tree) ⏭️

• 93ea780a contrib: aspect-oriented refactoring examples for crosscutting concerns (RE2026) #4719 — aspect-oriented refactoring examples (the old contrib/aop examples are not present in the current tree and should not be restored as a standalone stale example bundle without a fresh docs/architecture decision.) ⏭️

Click to expand

17b9d93c docs, 95344a44 refactor READMEs, 8dd4c4b1 remove review-bot config, 884747b8 merge commit, 1de30948 dep bump distroless, 0813c771 dep bump libloading, d4045117 merge artifact, 773a72ec merge commit, 01da1b4d remove RedNote badge, 40805d30 merge commit, a18fb885 dep bump cpal, 50425247 merge collision fix, be747242 web merge collision fix, 92440b2a docs only, e91806c1 trivial lint

---

How to help

1. Pick an unchecked item, verify whether the fix exists in the current crate structure
2. If missing: open a recovery PR porting the fix to the new architecture
3. Update the checkbox and note the PR number

Original commits can be viewed with git show <sha>. All were against the old monolithic src/ structure and need porting to crates/.

Attribution: Recovery PRs supersede the original author's merged work. Follow the superseded-PR attribution rules in docs/contributing/pr-discipline.md — include Co-authored-by trailers for the original author and list the original PR in the body.

[ilteoood]

@Audacity88 about ac355d40, I've opened a new PR for it and has been integrated as part of 0.73.0

[Audacity88]

@ilteoood confirmed — ac355d40 was re-merged as #5321 (9f245539f). Moved to "Verified re-applied" in the tracker.

Also batch-checked all 14 "needs investigation" items against current master:

Newly confirmed re-applied (11):

• 1e477f52 max_history_messages — config-driven capping in orchestrator/mod.rs:1211
• a993c96f stream_chat in turn_streamed — loop_.rs:547
• 3b9c184e Z.AI/Bailian alias — provider_aliases.rs:89
• e9c99f92 Matrix thread context — matrix.rs:256-292
• 3ce90480 path_prefix for paircode — gateway routes nested under prefix
• 516a9cfb temp dir canonicalization (macOS) — schema.rs:9418
• a279f5da dashboard Homebrew message — service/mod.rs:582
• 9ce1b67a duplicate tool call notifications — superseded by fix(agent): drop redundant narration push before AssistantToolCalls #6093
• 15a630e4 public bind error message — gateway/lib.rs:413
• 6044fa23 agent-browser Windows shim — browser.rs:247
• 59c9a2d1 claude-code test flakiness — superseded by fix(agent): drop redundant narration push before AssistantToolCalls #6093

Newly confirmed missing (2):

• 8c3c890c Mattermost HTTP timeout — build_channel_proxy_client_with_timeouts() exists but Mattermost still uses the non-timeout variant
• babe7f1e fix(heartbeat): inject current datetime into scheduler decision prompt #4554 scheduler datetime injection — build_decision_prompt() has no datetime context

Skipped (1):

• a50e2304 blob/main→blob/master link — extensions/movie/CONTRIBUTING_GUIDE.md doesn't exist in current tree

Tracker body updated to reflect.