finish
Skill2 starsChanged 9 months agoMITView on GitHub
Master pipeline — compile, review, fix, and finalize any document. Auto-detects format and runs the full quality loop.
22 other projects use a version of this.
Adopted by
Most recently taken up by
- MarvinJWendt/skillstheir own version, first seen Nov 8, 2026
- leinardi/adversarial-review-looptheir own version, first seen Nov 8, 2026
- MammadovEshgin/clean-code-skilltheir own version, first seen Nov 8, 2026
- hex/claude-sessionstheir own version, first seen Nov 3, 2026
- CG-5228/ClaudeCodeImpactLab_hackathontheir own version, first seen Nov 1, 2026
- danielcherubini/skillstheir own version, first seen Oct 25, 2026
Used with
Projects with it also use status, bootstrap and 10 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
AnthropicSkillstatus
Case status summary by audience — client-facing (plain language), internal (for the professor), or court-ready (formal caption format per local rules). Same facts, different framing and depth. Use when a student needs to update the client, brief the professor, or prepare a court status report.
Also in 5 projects that use this
OpenAISkillbootstrap
Project bootstrapping orchestrator for repos that depend on Vercel-linked resources (databases, auth, and managed integrations). Use when setting up or repairing a repository so linking, environment provisioning, env pulls, and first-run db/dev commands happen in the correct safe order.
Also in 3 projects that use this
AnthropicSkilldebug
Structured debugging session — reproduce, isolate, diagnose, and fix. Trigger with an error message or stack trace, "this works in staging but not prod", "something broke after the deploy", or when behavior diverges from expected and the cause isn't obvious.
Also in 5 projects that use this
Hugging FaceSkillimplement
Make tests pass. Invoke after /write-tests produces failing tests.
Also in 6 projects that use this
dillonluiSkilldesign-review
Review a view or flow for UX and UI issues. Provides severity-ranked, actionable findings grounded in established design heuristics. Use when the user asks for design feedback, UX review, or UI critique.
Also in 3 projects that use this
SentrySkillreview
Fixture review skill.
Also in 8 projects that use this
- +1 more from anthropics
Other versions22 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 22
- hacan359/tonkatsu_box★ 597same name
- alexeykrol/claude-code-starter★ 194same name
- hex/claude-sessions★ 54same name
- rlajous/claude-code-commands★ 30same name
- alexiolan/craft-skills★ 17same name
- danielcherubini/skills★ 14same name
- keefar/ticket-flow★ 4same name
- mmjclayton/agent-sop★ 4same name
- andreasbaumgartner/gilo★ 2same name
- leinardi/adversarial-review-loop★ 2same name
12 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository23 other files it gives its agents.
- compileSkill
- context-statusSkill
- create-documentSkill
- deploySkill
- devils-advocateSkill
- diagramSkill
- doc-reviewSkill
- excellenceSkill
- extract-diagramsSkill
- google-tasksSkill
- image-searchSkill
- learnSkill
- mindmapSkill
- pedagogy-reviewSkill
- proofreadSkill
- qaSkill
- sec-edgarSkill
- translateSkill
- typstSkill
- validate-bibSkill
- verifySkill
- visual-auditSkill
- waybackSkill
The file
We split this skill into the rules it gives an agent
10 rules. 0 are common, 9 are distinctive, and 1 is very specific to this project.
0
Common
Most of the 18 edited versions we read keep these.
9
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 52
Read each PNG with the Read tool
- Distinctive: line 66
If the visual spot-check caught simple issues, fix them now so review agents start from a clean baseline.
- Distinctive: line 68
Scope: Only fix issues that are:
- Distinctive: line 99
No new visual issues were introduced
- Distinctive: line 110
If typst-reviewer flags "content would benefit from visuals":
- Very specific: line 111
Call gemini-generate-image MCP for conceptual illustrations
- Distinctive: line 119
If score >= 80 AND zero CRITICAL issues → ready to commit
- Distinctive: line 120
If score < 80 → report remaining issues, do NOT auto-commit
- Distinctive: line 161
Always re-compile after fixes to verify they don't break anything
- Distinctive: line 162
Always re-render affected pages to PNG after fixes to confirm visual correctness
“Common” is measured: kept by at least 6 in 10 of the 18 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Finish Document
Master pipeline that compiles, reviews, fixes, and finalizes any supported document type.
Format Detection
Determine file type from extension:
.typ→ Typst document (slides, docs, guides, CVs).qmd→ Quarto RevealJS slides.py→ Python script
Note: For
.tex(LaTeX/Beamer), configure the finish pipeline at project level.
For .typ files, also detect document type from content:
- Contains
#import "@preview/polylux"or#import "@preview/touying"→ Typst slides - Contains
#import "@local/qk"→ Uses qk component library - Otherwise → Typst document
Pipeline Phases
Phase 0: Detect File Type
- Parse
$ARGUMENTS, resolve file path - Detect format from extension and content
Phase 1: Compile + Visual Spot-Check
Step 1 — Compile:
- Route to
/compilewith the detected format - Hard gate: exit code must be 0. Do not proceed until compilation succeeds.
Step 2 — Determine sample pages:
- Documents <=3 pages → all pages
- Documents >3 pages → page 1, middle page, last page
- Presentations <30 slides → all slides
- Presentations 30+ → first 3, middle 3, last 3
Step 3 — Render PNGs:
typst compile FILE.typ /tmp/finish-preview-{0p}.png --pages [SAMPLE]Step 4 — Visual inspection:
- Read each PNG with the Read tool
- Apply the full visual checklist from
typst/references/visual-verification.md:- Content overflow / cut off?
- Blank half-pages?
- Font fallback squares (missing font)?
- Missing images / broken references?
- Cramped text / unbalanced layout?
- Heading hierarchy consistent?
- Table headers repeating across pages?
Step 5 — Record issues for Phase 1.5 or Phase 3 fix list.
Phase 1.5: Quick Auto-Fix Before Review
If the visual spot-check caught simple issues, fix them now so review agents start from a clean baseline.
Scope: Only fix issues that are:
- Clearly identified in the PNG inspection
- Fixable without design decisions (overflow, missing alt text, font fallback)
Steps:
- Apply fixes:
- Overflow → split content, adjust spacing, or reduce font size
- Missing alt text → add generic alt descriptions
- Font fallback → switch to documented font stack (see typst skill)
- Blank half-pages → adjust
figure(placement: auto)or page breaks
- Re-compile:
typst compile FILE.typ - Re-render affected sample pages to PNG
- Visually verify fixes resolved the issue AND didn't introduce new ones
- Max 2 mini-rounds, then proceed to Phase 2 regardless
Phase 2: Parallel Review
- Route to
/doc-reviewwith the file - Collect all agent reports
Phase 3: Fix Critical Issues (max 5 rounds)
For each fix round:
- Apply fixes — priority order: Critical → Major → Minor
- Re-compile:
typst compile FILE.typ - Re-render affected pages to PNG:
typst compile FILE.typ /tmp/finish-fix-{0p}.png --pages [AFFECTED] - Visually verify that:
- Fixes resolved the flagged issue
- No new visual issues were introduced
- If new visual issues → add to fix list for next round
Format-specific notes:
For .qmd files:
- Launch quality-fixer agent with critic report
- Re-render after fixes
For .typ files:
- Apply fixes directly via Edit tool
- If typst-reviewer flags "content would benefit from visuals":
- Call
gemini-generate-imageMCP for conceptual illustrations - Invoke
/image-searchfor real photos/logos
- Call
For .py files:
- Apply fixes directly via Edit tool
- Re-run to verify output
Phase 4: Commit Decision
- If score >= 80 AND zero CRITICAL issues → ready to commit
- If score < 80 → report remaining issues, do NOT auto-commit
Phase 5: Summary Report
# Finish Report: [Filename]
## Pipeline Summary
| Phase | Status | Details |
|-------|--------|---------|
| Compile | PASS/FAIL | [output file, warnings] |
| Visual Spot-Check | PASS/[N issues] | [issues found and auto-fixed] |
| Review | [score] | [agent count, issue count] |
| Fix | [rounds] | [issues fixed / remaining] |
| Final Score | [score] | [EXCELLENT/GOOD/NEEDS WORK/POOR] |
## Score Progression
- Initial: [score] ([N] critical, [M] major)
- After auto-fix: [score] ([N] critical, [M] major)
- Final: [score] ([N] critical, [M] major)
- Rounds: [N]
## Fixes Applied
[List of changes made]
## Remaining Issues (if any)
[List issues that could not be auto-fixed]Iteration Limits
See typst/references/quality-gates.md for authoritative limits.
- Fix rounds: max 5
- Auto-fix mini-rounds (Phase 1.5): max 2
- Verification retries: max 2 per round
- Never loop indefinitely
Important
- For
.typfiles, visual tool invocation (gemini-generate-imageMCP,/image-search) only happens when the reviewer explicitly flags missing visuals - Always re-compile after fixes to verify they don't break anything
- Always re-render affected pages to PNG after fixes to confirm visual correctness
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

