Navigation index for the docs/ directory. Nothing here is load-bearing for a pipeline run — models read AGENTS.md, rules/, roles/, and skills/ only.
- comparative-analysis.md — Stagecraft's Claude/Codex lineage and its position across delivery control, spec systems, methodologies, context systems, and agent runtimes including Omnigent
- adoption-guide.md — pilot script, common objections, success criteria for the 2-week trial
- presentation-notes.md — slide deck + speaker notes for pitching to a team or stakeholder
- walkthroughs/soc2-evidence-collector.md — end-to-end showcase: building a SOC 2 evidence collector through the full 18-stage pipeline
- walkthroughs/stage-04-split-host.md — multi-host contract stress-test trace
- user-guide.md — daily-use reference: running stages, multi-host setups, headless mode
- user-guide.md § Prototype first, harden later — short prototype-mode tutorial using a token cost estimator example
- tracks.md — which of the six tracks to pick for a given change, plus why prototype mode is a pre-SDLC workflow rather than a track
- conventions.md — pipeline markers operators read and write (
QUESTION:,BLOCKER:, magic comments) - runbooks/README.md — troubleshooting index: symptom → runbook section
- cost.md — cost tracking, pricing table, and budget workflow
- faq.md — operational questions and common gotchas
- git-workflow.md — branch setup and PR timing for a Stagecraft pipeline run
- ci.md — GitHub Actions workflow template and environment variables
- reference/environment-variables.md — all environment variables Stagecraft reads, grouped by subsystem
- memory.md — persistent project memory: embedder options and shared store
- observability.md — OpenTelemetry span schema and collector setup
- evidence.md — offline readiness, explicit resolution acceptance, and privacy-safe export
- reproducibility.md — gate fingerprint fields, replay readiness, drift detection
- omnigent-runtime.md — Omnigent as a Stagecraft host runtime: adapter boundary, shipped slice, and phased policy/session bridge
- runbooks/escalation.md — what to read and decide when
devteam nextsaysresolve-escalation - runbooks/fix-and-retry.md — fix-and-retry halt: red-team, QA, pre-review, peer-review (11 cases)
- runbooks/open-followups.md — extract ticket stubs from
open_followups[]after a run - runbooks/deploy-failure.md — Stage 8 failure: classify, diagnose, rollback, retry
- runbooks/autonomous-run.md —
devteam run: bounded autonomous driver reference - ../hosts/docker/README.md — Docker-based headless runner for unattended local orchestration
- runbooks/repair-flow.md —
devteam run --repair: diagnosis gate approval, scope-gate FAIL recovery, tri-state reproduction - runbook-template.md — template and guide for writing new runbooks
- guides/dogfooding.md — running Stagecraft against its own source tree: one-time setup, per-feature workflow, budget guidance
- concepts.md — six pipeline primitives (stage, role, workstream, host, gate, track) in one table
- methodology.md — ATDD loop, phase-gate progression, adversarial layer, four coding principles
- FEATURES.md — every shipped feature organized by area
- BACKLOG.md — prioritized open ideas by bucket
- TESTING.md — test structure and guidance
- adr/README.md — architecture decision records index
- adr/011-authenticated-gate-chain.md — HMAC gate authentication, strict policy, and trust boundaries
- adr/012-explicit-resolution-acceptance.md — hash-bound human acceptance evidence for H3
- adr/013-openai-compat-shell-execution.md — openai-compat shell execution: native bash tool, first-class host status
- adr/014-docker-headless-runner.md — Docker runner trust boundary and packaging decision
- autonomous-execution-design.md — design notes for the bounded autonomous driver (companion to ADR-003)
- openai-compat-design.md — options analysis and design notes for the openai-compat host adapter (companion to ADR-013)
- ../plans/phase-24-omnigent-runtime.md — phased plan for Omnigent runtime integration
- spec-authoring.md — writing AC-N criteria and scaffolding the spec file
- migration-safety.md — veto criteria, gate fields, and the migration heuristic
- red-team.md — 10 attack surfaces, gate fields, routing, and how it differs from security review
- verification-beyond-tests.md — property-based, mutation, and formal verification: candidates and skip policy
- brief-template.md — section-by-section guide to filling out
pipeline/brief.md - design-spec-template.md — section-by-section guide to filling out
pipeline/design-spec.md - GAP-ANALYSIS.md — archived gap analysis (redirect); active gaps tracked in BACKLOG.md
Models read AGENTS.md + rules/ + roles/ + skills/ only. Nothing in this directory is loaded during a pipeline run.