Files
msd-core/gsd-core/workflows/extract-learnings.md
Tom Boucher 7c07fce70f fix(#381): make gsd_run launcher reachable in fresh-shell-per-block runtimes (#1084)
* fix(#381): make gsd_run launcher reachable in fresh-shell-per-block runtimes

On runtimes that execute each fenced bash block in a separate shell process
(e.g. Claude Code — documented behavior: each Bash command is a separate
process; inline shell functions and exported vars do not persist between
calls), the once-per-file gsd_run() function was undefined in every block
after the preamble block, and the call was swallowed by
`2>/dev/null || echo "{}"` into silent empty state.

Fix (budget-neutral session-level resolution):
- Ship gsd-core/bin/gsd_run, a POSIX sh wrapper that symlink-resolves its own
  location and execs the co-located gsd-tools.cjs. Exposed on PATH via the npm
  `bin` field (global installs) and shipped to local installs via the recursive
  gsd-core/ copy.
- The per-file launcher preamble now appends `export PATH='<bindir>':"$PATH"`
  to the file named by $CLAUDE_ENV_FILE (Claude Code's documented
  env-persistence mechanism) so later fresh-shell blocks resolve gsd_run from
  PATH. Guarded as a strict no-op when CLAUDE_ENV_FILE is unset; the inline
  gsd_run() definition remains the fallback for all other runtimes. The
  single-quoted dir neutralizes shell metacharacters at source time.
- Propagated via scripts/sync-runtime-launcher.cjs to all launcher-using files.
- XL workflow byte budget 93000 -> 93200 (the ~130B clause pushes plan-phase.md
  to 93135; legitimate content growth, ratchet-up per #717).

Regression tests (I)/(J) in runtime-launcher-parity.test.cjs cover wrapper
delegation and end-to-end PATH persistence (sourcing the env file with a
space-bearing install path).

Known limitation: an install path containing a literal single-quote yields a
malformed env-file line and falls back to the status quo (no regression);
rare on sanitized home directories.

Closes #381

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

* docs(#381): add changeset for gsd_run fresh-shell reachability fix

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

* test(#381): scope test (J) bare-PATH execution to POSIX (Windows Git Bash exec bit)

Windows Git Bash (msys2) does not honor Node's chmod exec bit for
PATH-executing extension-less scripts, so the bare `gsd_run` command lookup
failed there even though the env-file PATH persistence was correct. The
env-file content assertions (the fix's actual cross-platform logic) still run
on every platform; only the final source-and-execute sub-step is gated to
non-win32. Global installs on Windows are covered by npm's generated bin shim.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

---------

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-11 21:42:52 -04:00

12 KiB

Extract decisions, lessons learned, patterns discovered, and surprises encountered from completed phase artifacts into a structured LEARNINGS.md file. Captures institutional knowledge that would otherwise be lost between phases.

<required_reading> Read all files referenced by the invoking prompt's execution_context before starting. </required_reading>

Analyze completed phase artifacts (PLAN.md, SUMMARY.md, VERIFICATION.md, UAT.md, STATE.md) and extract structured learnings into 4 categories: decisions, lessons, patterns, and surprises. Each extracted item includes source attribution. The output is a LEARNINGS.md file with YAML frontmatter containing metadata about the extraction. Parse arguments and load project state:
_GSD_SHIM_NAME="gsd-tools.cjs"; _GSD_RUNTIME_ROOT="${RUNTIME_DIR:-$(git rev-parse --show-toplevel 2>/dev/null || pwd)}"; GSD_TOOLS="${_GSD_RUNTIME_ROOT}/gsd-core/bin/${_GSD_SHIM_NAME}"; if [ -f "$GSD_TOOLS" ]; then gsd_run() { node "$GSD_TOOLS" "$@"; }; elif [ -f "${_GSD_RUNTIME_ROOT}/.claude/gsd-core/bin/${_GSD_SHIM_NAME}" ]; then GSD_TOOLS="${_GSD_RUNTIME_ROOT}/.claude/gsd-core/bin/${_GSD_SHIM_NAME}"; gsd_run() { node "$GSD_TOOLS" "$@"; }; elif [ -f "${_GSD_RUNTIME_ROOT}/.codex/gsd-core/bin/${_GSD_SHIM_NAME}" ]; then GSD_TOOLS="${_GSD_RUNTIME_ROOT}/.codex/gsd-core/bin/${_GSD_SHIM_NAME}"; gsd_run() { node "$GSD_TOOLS" "$@"; }; elif command -v gsd-tools >/dev/null 2>&1; then GSD_TOOLS="$(command -v gsd-tools)"; gsd_run() { "$GSD_TOOLS" "$@"; }; elif [ -f "$HOME/.claude/gsd-core/bin/${_GSD_SHIM_NAME}" ]; then GSD_TOOLS="$HOME/.claude/gsd-core/bin/${_GSD_SHIM_NAME}"; gsd_run() { node "$GSD_TOOLS" "$@"; }; elif [ -f "${HERMES_HOME:-$HOME/.hermes}/gsd-core/bin/${_GSD_SHIM_NAME}" ]; then GSD_TOOLS="${HERMES_HOME:-$HOME/.hermes}/gsd-core/bin/${_GSD_SHIM_NAME}"; gsd_run() { node "$GSD_TOOLS" "$@"; }; elif [ -f "${CURSOR_CONFIG_DIR:-$HOME/.cursor}/gsd-core/bin/${_GSD_SHIM_NAME}" ]; then GSD_TOOLS="${CURSOR_CONFIG_DIR:-$HOME/.cursor}/gsd-core/bin/${_GSD_SHIM_NAME}"; gsd_run() { node "$GSD_TOOLS" "$@"; }; elif [ -f "${CODEX_HOME:-$HOME/.codex}/gsd-core/bin/${_GSD_SHIM_NAME}" ]; then GSD_TOOLS="${CODEX_HOME:-$HOME/.codex}/gsd-core/bin/${_GSD_SHIM_NAME}"; gsd_run() { node "$GSD_TOOLS" "$@"; }; elif [ -f "${GEMINI_CONFIG_DIR:-$HOME/.gemini}/gsd-core/bin/${_GSD_SHIM_NAME}" ]; then GSD_TOOLS="${GEMINI_CONFIG_DIR:-$HOME/.gemini}/gsd-core/bin/${_GSD_SHIM_NAME}"; gsd_run() { node "$GSD_TOOLS" "$@"; }; elif [ -f "${COPILOT_CONFIG_DIR:-$HOME/.copilot}/gsd-core/bin/${_GSD_SHIM_NAME}" ]; then GSD_TOOLS="${COPILOT_CONFIG_DIR:-$HOME/.copilot}/gsd-core/bin/${_GSD_SHIM_NAME}"; gsd_run() { node "$GSD_TOOLS" "$@"; }; elif [ -f "${WINDSURF_CONFIG_DIR:-$HOME/.codeium/windsurf}/gsd-core/bin/${_GSD_SHIM_NAME}" ]; then GSD_TOOLS="${WINDSURF_CONFIG_DIR:-$HOME/.codeium/windsurf}/gsd-core/bin/${_GSD_SHIM_NAME}"; gsd_run() { node "$GSD_TOOLS" "$@"; }; elif [ -f "${AUGMENT_CONFIG_DIR:-$HOME/.augment}/gsd-core/bin/${_GSD_SHIM_NAME}" ]; then GSD_TOOLS="${AUGMENT_CONFIG_DIR:-$HOME/.augment}/gsd-core/bin/${_GSD_SHIM_NAME}"; gsd_run() { node "$GSD_TOOLS" "$@"; }; elif [ -f "${TRAE_CONFIG_DIR:-$HOME/.trae}/gsd-core/bin/${_GSD_SHIM_NAME}" ]; then GSD_TOOLS="${TRAE_CONFIG_DIR:-$HOME/.trae}/gsd-core/bin/${_GSD_SHIM_NAME}"; gsd_run() { node "$GSD_TOOLS" "$@"; }; elif [ -f "${QWEN_CONFIG_DIR:-$HOME/.qwen}/gsd-core/bin/${_GSD_SHIM_NAME}" ]; then GSD_TOOLS="${QWEN_CONFIG_DIR:-$HOME/.qwen}/gsd-core/bin/${_GSD_SHIM_NAME}"; gsd_run() { node "$GSD_TOOLS" "$@"; }; elif [ -f "${CODEBUDDY_CONFIG_DIR:-$HOME/.codebuddy}/gsd-core/bin/${_GSD_SHIM_NAME}" ]; then GSD_TOOLS="${CODEBUDDY_CONFIG_DIR:-$HOME/.codebuddy}/gsd-core/bin/${_GSD_SHIM_NAME}"; gsd_run() { node "$GSD_TOOLS" "$@"; }; elif [ -f "${CLINE_CONFIG_DIR:-$HOME/.cline}/gsd-core/bin/${_GSD_SHIM_NAME}" ]; then GSD_TOOLS="${CLINE_CONFIG_DIR:-$HOME/.cline}/gsd-core/bin/${_GSD_SHIM_NAME}"; gsd_run() { node "$GSD_TOOLS" "$@"; }; elif [ -f "${GROK_AGENTS_HOME:-$HOME/.agents}/gsd-core/bin/${_GSD_SHIM_NAME}" ]; then GSD_TOOLS="${GROK_AGENTS_HOME:-$HOME/.agents}/gsd-core/bin/${_GSD_SHIM_NAME}"; gsd_run() { node "$GSD_TOOLS" "$@"; }; elif [ -f "${ANTIGRAVITY_CONFIG_DIR:-$HOME/.gemini/antigravity}/gsd-core/bin/${_GSD_SHIM_NAME}" ]; then GSD_TOOLS="${ANTIGRAVITY_CONFIG_DIR:-$HOME/.gemini/antigravity}/gsd-core/bin/${_GSD_SHIM_NAME}"; gsd_run() { node "$GSD_TOOLS" "$@"; }; elif [ -f "${OPENCODE_CONFIG_DIR:-${XDG_CONFIG_HOME:-$HOME/.config}/opencode}/gsd-core/bin/${_GSD_SHIM_NAME}" ]; then GSD_TOOLS="${OPENCODE_CONFIG_DIR:-${XDG_CONFIG_HOME:-$HOME/.config}/opencode}/gsd-core/bin/${_GSD_SHIM_NAME}"; gsd_run() { node "$GSD_TOOLS" "$@"; }; elif [ -f "${KILO_CONFIG_DIR:-${XDG_CONFIG_HOME:-$HOME/.config}/kilo}/gsd-core/bin/${_GSD_SHIM_NAME}" ]; then GSD_TOOLS="${KILO_CONFIG_DIR:-${XDG_CONFIG_HOME:-$HOME/.config}/kilo}/gsd-core/bin/${_GSD_SHIM_NAME}"; gsd_run() { node "$GSD_TOOLS" "$@"; }; else echo "ERROR: gsd-tools.cjs not found at $GSD_TOOLS and gsd-tools is not on PATH. Run: npx -y @opengsd/gsd-core@latest --claude --local" >&2; exit 1; fi; if [ -n "${CLAUDE_ENV_FILE:-}" ] && [ -n "${GSD_TOOLS:-}" ]; then printf "export PATH='%s':\"\$PATH\"\n" "${GSD_TOOLS%/*}" >> "$CLAUDE_ENV_FILE" 2>/dev/null || true; fi
INIT=$(gsd_run query init.phase-op "${PHASE_ARG}")
if [[ "$INIT" == @file:* ]]; then INIT=$(cat "${INIT#@file:}"); fi

Parse from init JSON: phase_found, phase_dir, phase_number, phase_name, padded_phase.

If phase not found, exit with error: "Phase {PHASE_ARG} not found."

Read the phase artifacts. PLAN.md and SUMMARY.md are required; VERIFICATION.md, UAT.md, and STATE.md are optional.

Required artifacts:

  • ${PHASE_DIR}/*-PLAN.md — all plan files for the phase
  • ${PHASE_DIR}/*-SUMMARY.md — all summary files for the phase

If PLAN.md or SUMMARY.md files are not found or missing, exit with error: "Required artifacts missing. PLAN.md and SUMMARY.md are required for learning extraction."

Optional artifacts (read if available, skip if not found):

  • ${PHASE_DIR}/*-VERIFICATION.md — verification results
  • ${PHASE_DIR}/*-UAT.md — user acceptance test results
  • .planning/STATE.md — project state with decisions and blockers

Track which optional artifacts are missing for the missing_artifacts frontmatter field.

Analyze all collected artifacts and extract learnings into 4 categories:

1. Decisions

Technical and architectural decisions made during the phase. Look for:

  • Explicit decisions documented in PLAN.md or SUMMARY.md
  • Technology choices and their rationale
  • Trade-offs that were evaluated
  • Design decisions recorded in STATE.md

Each decision entry must include:

  • What was decided
  • Why it was decided (rationale)
  • Source: attribution to the artifact where the decision was found (e.g., "Source: 03-01-PLAN.md")

2. Lessons

Things learned during execution that were not known beforehand. Look for:

  • Unexpected complexity in SUMMARY.md
  • Issues discovered during verification in VERIFICATION.md
  • Failed approaches documented in SUMMARY.md
  • UAT feedback that revealed gaps

Each lesson entry must include:

  • What was learned
  • Context for the lesson
  • Source: attribution to the originating artifact

3. Patterns

Reusable patterns, approaches, or techniques discovered. Look for:

  • Successful implementation patterns in SUMMARY.md
  • Testing patterns from VERIFICATION.md or UAT.md
  • Workflow patterns that worked well
  • Code organization patterns from PLAN.md

Each pattern entry must include:

  • Pattern name/description
  • When to use it
  • Source: attribution to the originating artifact

4. Surprises

Unexpected findings, behaviors, or outcomes. Look for:

  • Things that took longer or shorter than estimated
  • Unexpected dependencies or interactions
  • Edge cases not anticipated in planning
  • Performance or behavior that differed from expectations

Each surprise entry must include:

  • What was surprising
  • Impact of the surprise
  • Source: attribution to the originating artifact
**What this step is:** `capture_thought` is an **optional convention**, not a bundled GSD tool. GSD does not ship one and does not require one. The step is a hook for users who run a memory / knowledge-base MCP server (for example ExoCortex-style servers, `claude-mem`, or `mem0`-style servers) that exposes a tool with this exact name. If any MCP server in the current session provides a `capture_thought` tool with the signature below, each extracted learning is routed through it with metadata. If no such tool is present, the step is a silent no-op — `LEARNINGS.md` is always the primary output.

Detection: Check whether a tool named capture_thought is available in the current session. Do not assume any specific MCP server is connected.

If available, call once per extracted learning:

capture_thought({
  category: "decision" | "lesson" | "pattern" | "surprise",
  phase: PHASE_NUMBER,
  content: LEARNING_TEXT,
  source: ARTIFACT_NAME
})

If not available (no MCP server in the session exposes this tool, or the runtime does not support it), skip the step silently and continue. The workflow must not fail or warn — this is expected behavior for users who do not run a knowledge-base MCP.

Write the LEARNINGS.md file to the phase directory. If a previous LEARNINGS.md exists, overwrite it (replace the file entirely).

Output path: ${PHASE_DIR}/${PADDED_PHASE}-LEARNINGS.md

The file must have YAML frontmatter with these fields:

---
phase: {PHASE_NUMBER}
phase_name: "{PHASE_NAME}"
project: "{PROJECT_NAME}"
generated: "{ISO_DATE}"
counts:
  decisions: {N}
  lessons: {N}
  patterns: {N}
  surprises: {N}
missing_artifacts:
  - "{ARTIFACT_NAME}"
---

Individual items may carry an optional graduated: annotation (added by graduation.md when a cluster is promoted):

**Graduated:** {target-file}:{ISO_DATE}

This annotation is appended after the item's existing fields and prevents the item from being re-surfaced in future graduation scans. Do not add this field during extraction — it is written only by the graduation workflow.

The body follows this structure:

# Phase {PHASE_NUMBER} Learnings: {PHASE_NAME}

## Decisions

### {Decision Title}
{What was decided}

**Rationale:** {Why}
**Source:** {artifact file}

---

## Lessons

### {Lesson Title}
{What was learned}

**Context:** {context}
**Source:** {artifact file}

---

## Patterns

### {Pattern Name}
{Description}

**When to use:** {applicability}
**Source:** {artifact file}

---

## Surprises

### {Surprise Title}
{What was surprising}

**Impact:** {impact description}
**Source:** {artifact file}
Update STATE.md to reflect the learning extraction:
gsd_run query state.update "Last Activity" "$(date +%Y-%m-%d)"
``` ---------------------------------------------------------------

Learnings Extracted: Phase {X} — {Name}

Decisions: {N} Lessons: {N} Patterns: {N} Surprises: {N} Total: {N}

Output: {PHASE_DIR}/{PADDED_PHASE}-LEARNINGS.md

Missing artifacts: {list or "none"}

Next steps:

  • Review extracted learnings for accuracy
  • /gsd:progress — see overall project state
  • /gsd:execute-phase {next} — continue to next phase

</step>

</process>

<success_criteria>
- [ ] Phase artifacts located and read successfully
- [ ] All 4 categories extracted: decisions, lessons, patterns, surprises
- [ ] Each extracted item has source attribution
- [ ] LEARNINGS.md written with correct YAML frontmatter
- [ ] Missing optional artifacts tracked in frontmatter
- [ ] capture_thought integration attempted if tool available
- [ ] STATE.md updated with extraction activity
- [ ] User receives summary report
</success_criteria>

<critical_rules>
- PLAN.md and SUMMARY.md are required — exit with clear error if missing
- VERIFICATION.md, UAT.md, and STATE.md are optional — extract from them if present, skip gracefully if not found
- Every extracted learning must have source attribution back to the originating artifact
- Running extract-learnings twice on the same phase must overwrite (replace) the previous LEARNINGS.md, not append
- Do not fabricate learnings — only extract what is explicitly documented in artifacts
- If capture_thought is unavailable, the workflow must not fail — graceful degradation to file-only output
- LEARNINGS.md frontmatter must include counts for all 4 categories and list any missing_artifacts
</critical_rules>