fix: harden generated Graphify instructions

[andrebrait]

Summary

• pass INPUT_PATH and SPEC_PATH to generated Python snippets as argv instead of embedding them in Python source
• preserve argv when translating POSIX snippets to PowerShell here-strings
• use host-neutral subagent terminology and source transcription hints only from prior analysis
• regenerate and bless all 134 platform artifacts

Verification

• uv run pytest tests/test_skillgen.py -q (67 passed)
• uv run python -m tools.skillgen --check
• uv run python -m tools.skillgen --monolith-roundtrip
• apostrophe-bearing path argv smoke test

A full uv run pytest -q reached 5,417 passes / 95 skips with 6 unrelated failures: four missing optional openai dependency tests, the pre-existing hidden-fixture path assertion (the worktree path itself is hidden), and a timing-sensitive TypeScript scaling assertion.

[andrebrait]

@coderabbitai review

[Copilot]

🟢 Approval recommended

The generator changes are cohesive, the PowerShell translation is updated to preserve argv tails, and the new/updated tests directly cover the intended hardening behavior.

Pull request overview

This PR hardens Graphify’s generated skill/runbook instructions by ensuring user-provided paths (e.g., INPUT_PATH, SPEC_PATH) are passed to embedded Python snippets via argv rather than being embedded inside Python source, and updates the PowerShell translation logic so those argv tails survive when converting POSIX python -c blocks into PowerShell here-strings. It also refreshes transcription guidance to rely on prior analysis artifacts (when available) and regenerates the derived platform artifacts accordingly.

Changes:

• Pass INPUT_PATH / SPEC_PATH to generated Python snippets as command-line arguments (sys.argv[...]) instead of embedding placeholder strings in Python code.
• Update the POSIX→PowerShell bash-block translator to recognize and preserve Python-invocation “close” lines that include argv tails.
• Regenerate and update platform fragments/expected outputs, plus add regression tests covering argv preservation and removal of quoted placeholders.

File summaries

File	Description
tools/skillgen/gen.py	Adds PowerShell close-line translations for argv-bearing Python invocations and sanctions new monolith diff lines.
tools/skillgen/fragments/references/shared/update.md	Updates shared update reference snippets to read INPUT_PATH from argv and pass it via shell args.
tools/skillgen/fragments/references/shared/transcribe.md	Revises transcription strategy text to prefer prior analysis (when present).
tools/skillgen/fragments/core/devin.md	Updates devin monolith instructions to pass INPUT_PATH via argv and adjusts transcription guidance.
tools/skillgen/fragments/core/aider.md	Updates aider monolith instructions to pass INPUT_PATH via argv and adjusts transcription guidance.
tools/skillgen/expected/graphify__skills__windows__references__update.md	Updates expected Windows update reference output to argv-based INPUT_PATH.
tools/skillgen/expected/graphify__skills__windows__references__transcribe.md	Updates expected Windows transcription reference output to “previous analysis” guidance.
tools/skillgen/expected/graphify__skills__vscode__references__update.md	Updates expected VS Code update reference output to argv-based INPUT_PATH.
tools/skillgen/expected/graphify__skills__vscode__references__transcribe.md	Updates expected VS Code transcription reference output to “previous analysis” guidance.
tools/skillgen/expected/graphify__skills__trae__references__update.md	Updates expected Trae update reference output to argv-based INPUT_PATH.
tools/skillgen/expected/graphify__skills__trae__references__transcribe.md	Updates expected Trae transcription reference output to “previous analysis” guidance.
tools/skillgen/expected/graphify__skills__pi__references__update.md	Updates expected Pi update reference output to argv-based INPUT_PATH.
tools/skillgen/expected/graphify__skills__pi__references__transcribe.md	Updates expected Pi transcription reference output to “previous analysis” guidance.
tools/skillgen/expected/graphify__skills__opencode__references__update.md	Updates expected OpenCode update reference output to argv-based INPUT_PATH.
tools/skillgen/expected/graphify__skills__opencode__references__transcribe.md	Updates expected OpenCode transcription reference output to “previous analysis” guidance.
tools/skillgen/expected/graphify__skills__kiro__references__update.md	Updates expected Kiro update reference output to argv-based INPUT_PATH.
tools/skillgen/expected/graphify__skills__kiro__references__transcribe.md	Updates expected Kiro transcription reference output to “previous analysis” guidance.
tools/skillgen/expected/graphify__skills__kilo__references__update.md	Updates expected Kilo update reference output to argv-based INPUT_PATH.
tools/skillgen/expected/graphify__skills__kilo__references__transcribe.md	Updates expected Kilo transcription reference output to “previous analysis” guidance.
tools/skillgen/expected/graphify__skills__droid__references__update.md	Updates expected Droid update reference output to argv-based INPUT_PATH.
tools/skillgen/expected/graphify__skills__droid__references__transcribe.md	Updates expected Droid transcription reference output to “previous analysis” guidance.
tools/skillgen/expected/graphify__skills__copilot__references__update.md	Updates expected Copilot update reference output to argv-based INPUT_PATH.
tools/skillgen/expected/graphify__skills__copilot__references__transcribe.md	Updates expected Copilot transcription reference output to “previous analysis” guidance.
tools/skillgen/expected/graphify__skills__codex__references__update.md	Updates expected Codex update reference output to argv-based INPUT_PATH.
tools/skillgen/expected/graphify__skills__codex__references__transcribe.md	Updates expected Codex transcription reference output to “previous analysis” guidance.
tools/skillgen/expected/graphify__skills__claw__references__update.md	Updates expected Claw update reference output to argv-based INPUT_PATH.
tools/skillgen/expected/graphify__skills__claw__references__transcribe.md	Updates expected Claw transcription reference output to “previous analysis” guidance.
tools/skillgen/expected/graphify__skills__claude__references__update.md	Updates expected Claude update reference output to argv-based INPUT_PATH.
tools/skillgen/expected/graphify__skills__claude__references__transcribe.md	Updates expected Claude transcription reference output to “previous analysis” guidance.
tools/skillgen/expected/graphify__skills__amp__references__update.md	Updates expected Amp update reference output to argv-based INPUT_PATH.
tools/skillgen/expected/graphify__skills__amp__references__transcribe.md	Updates expected Amp transcription reference output to “previous analysis” guidance.
tools/skillgen/expected/graphify__skills__agents__references__update.md	Updates expected Agents update reference output to argv-based INPUT_PATH.
tools/skillgen/expected/graphify__skills__agents__references__transcribe.md	Updates expected Agents transcription reference output to “previous analysis” guidance.
tools/skillgen/expected/graphify__skill-devin.md	Updates expected devin monolith output for argv-based path passing and transcription guidance.
tools/skillgen/expected/graphify__skill-aider.md	Updates expected aider monolith output for argv-based path passing and transcription guidance.
tests/test_skillgen.py	Adds coverage ensuring placeholders aren’t embedded in Python source, argv is preserved in PS translation, and transcription wording is updated.
graphify/skills/windows/references/update.md	Updates Windows update reference artifact to argv-based INPUT_PATH.
graphify/skills/windows/references/transcribe.md	Updates Windows transcription reference artifact to “previous analysis” guidance.
graphify/skills/vscode/references/update.md	Updates VS Code update reference artifact to argv-based INPUT_PATH.
graphify/skills/vscode/references/transcribe.md	Updates VS Code transcription reference artifact to “previous analysis” guidance.
graphify/skills/trae/references/update.md	Updates Trae update reference artifact to argv-based INPUT_PATH.
graphify/skills/trae/references/transcribe.md	Updates Trae transcription reference artifact to “previous analysis” guidance.
graphify/skills/pi/references/update.md	Updates Pi update reference artifact to argv-based INPUT_PATH.
graphify/skills/pi/references/transcribe.md	Updates Pi transcription reference artifact to “previous analysis” guidance.
graphify/skills/opencode/references/update.md	Updates OpenCode update reference artifact to argv-based INPUT_PATH.
graphify/skills/opencode/references/transcribe.md	Updates OpenCode transcription reference artifact to “previous analysis” guidance.
graphify/skills/kiro/references/update.md	Updates Kiro update reference artifact to argv-based INPUT_PATH.
graphify/skills/kiro/references/transcribe.md	Updates Kiro transcription reference artifact to “previous analysis” guidance.
graphify/skills/kilo/references/update.md	Updates Kilo update reference artifact to argv-based INPUT_PATH.
graphify/skills/kilo/references/transcribe.md	Updates Kilo transcription reference artifact to “previous analysis” guidance.
graphify/skills/droid/references/update.md	Updates Droid update reference artifact to argv-based INPUT_PATH.
graphify/skills/droid/references/transcribe.md	Updates Droid transcription reference artifact to “previous analysis” guidance.
graphify/skills/copilot/references/update.md	Updates Copilot update reference artifact to argv-based INPUT_PATH.
graphify/skills/copilot/references/transcribe.md	Updates Copilot transcription reference artifact to “previous analysis” guidance.
graphify/skills/codex/references/update.md	Updates Codex update reference artifact to argv-based INPUT_PATH.
graphify/skills/codex/references/transcribe.md	Updates Codex transcription reference artifact to “previous analysis” guidance.
graphify/skills/claw/references/update.md	Updates Claw update reference artifact to argv-based INPUT_PATH.
graphify/skills/claw/references/transcribe.md	Updates Claw transcription reference artifact to “previous analysis” guidance.
graphify/skills/claude/references/update.md	Updates Claude update reference artifact to argv-based INPUT_PATH.
graphify/skills/claude/references/transcribe.md	Updates Claude transcription reference artifact to “previous analysis” guidance.
graphify/skills/amp/references/update.md	Updates Amp update reference artifact to argv-based INPUT_PATH.
graphify/skills/amp/references/transcribe.md	Updates Amp transcription reference artifact to “previous analysis” guidance.
graphify/skills/agents/references/update.md	Updates Agents update reference artifact to argv-based INPUT_PATH.
graphify/skills/agents/references/transcribe.md	Updates Agents transcription reference artifact to “previous analysis” guidance.
graphify/skill.md	Updates primary skill instructions to argv-based path passing, PowerShell-neutral terminology (“subagent”), and prompt-file argv passing.
graphify/skill-vscode.md	Updates VS Code skill artifact to argv-based path passing and subagent terminology.
graphify/skill-trae.md	Updates Trae skill artifact to argv-based path passing and subagent terminology.
graphify/skill-pi.md	Updates Pi skill artifact to argv-based path passing and subagent terminology.
graphify/skill-opencode.md	Updates OpenCode skill artifact to argv-based path passing and subagent terminology.
graphify/skill-kilo.md	Updates Kilo skill artifact to argv-based path passing and subagent terminology.
graphify/skill-droid.md	Updates Droid skill artifact to argv-based path passing and subagent terminology.
graphify/skill-devin.md	Updates devin skill artifact to argv-based path passing and revised transcription wording.
graphify/skill-claw.md	Updates Claw skill artifact to argv-based path passing and subagent terminology.
graphify/skill-amp.md	Updates Amp skill artifact to argv-based path passing and subagent terminology.
graphify/skill-aider.md	Updates aider skill artifact to argv-based path passing and revised transcription wording.

Review details

• Files reviewed: 95/95 changed files
• Comments generated: 0
• Review effort level: Lite

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[andrebrait]

@coderabbitai review

[andrebrait]

CodeRabbit did not respond to the initial request or the one permitted retry; the required independent adversarial review and Copilot review completed cleanly on bc043e4.

[andrebrait]

Landed bc043e4 by signed fast-forward to the integration branch; closing the now-empty PR.