autoplan
garrytan/gstack / autoplan
runs CEO, design, engineering and DX reviews , which improves: faster plan reviews, consistent decisions
Details
- judge’s score
- 5/10
- place
- outside the packs
- publisher
- garrytan
- status
- new
- pinned commit
- c580561 on GitHub (new tab)
- skill’s last commit
- Oct 20, 2026
- repo stars
- 137,030
- main type
- Building with Claude
Judgement
SKILL.md truncated (24k/64k) and appendices truncated at 5k. Meta-review pipeline specific to gstack, with no verifiable procedure visible. Clear triggering, but heavy dependencies on other skills and absent tools.
reservations
- SKILL.md truncated at 24,000 characters out of 64,740: the rest (steps, decision principles) was not assessed
- Appendices ceo, design, dx and eng truncated at 5,000 characters; manifest.json, DRIVER.md and the bin/ files not read
- Depends on gstack skills and scripts (gstack-skill-start, gstack-codex-probe, snapshot tool) that are not provided
- Automatically generated file, with a very long meta-protocol preamble for little skill-specific content
- Triggering
- 1/2
- Added value
- 1/2
- Precision
- 1/2
- Scope
- 1/2
- Quality
- 1/2
Alerts
7 alerts, by category
Static scan
- critical risk
frontmatter hooks (active for the whole session)
- high risk
bin/ folder
- high risk
scripts run by the SKILL.md
bin/phase-publication-hook
- info risk
allowed-tools frontmatter
Bash - Read - Write - Edit - Glob - Grep - WebSearch - AskUserQuestion
- info risk
bundled scripts, not run by the SKILL.md
bin/guard-journal.ts, bin/guard-log.ts, bin/guard-reasons.ts, bin/owned-read.ts, bin/phase-publication-hook.ts
Execution
- high risk
Decisions are made automatically without human validation, including on scope points and assumptions ("never stop mid-pipeline")
Instruction
- medium risk
The skill sends the full plan and source code to an external engine (Codex) via a shell command, without visible user confirmation
Original description
Auto-review pipeline — reads the full CEO, design, eng, and DX review skills from disk and runs them sequentially with auto-decisions using 6 decision principles. (gstack)
added on Oct 19, 2026 · license MIT