overbaked
Skill82 starsChanged 42 days agoMITView on GitHub
Audit a document, plan, or code for over-engineering, verbosity, and scope creep. Use when you've just written something and want a point-by-point check before finalizing.
In this repository27 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- ai-reviewSkill
- brieflySkill
- commit-batchingSkill
- commit-batchSkill
- create-alertSkill
- disclose-projectSkill
- discloseSkill
- evolve-claude-mdSkill
- figure-reviewSkill
- grillingSkill
- grill-meSkill
- handoffSkill
- lit-reviewSkill
- pathfinderSkill
- pickupSkill
- recover-sessionsSkill
- repo-initSkill
- retroSkill
- reviewer-2Skill
- slack-messageSkill
- spot-aiSkill
- tab-setupSkill
- unstaleSkill
- worklogSkill
- write-new-skillSkill
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 4 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 47
If something is fine, skip it — do not manufacture findings.
- Distinctive: line 100
If either is missing, ask for both in one question.
- Distinctive: line 101
If the material is provided but the audience isn't, infer it from the material's tone, jargon, and assumed knowledge — then state your …
- Distinctive: line 103
Use the audience to calibrate every finding.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
You are a ruthless editor. Your job is to find every place where the target material does more work than the task requires — and propose a tighter alternative. Do not soften findings. Be specific and direct.
What counts as overbaked
Verbosity
- Filler phrases ("it is worth noting", "as mentioned above", "in order to")
- Restating what was just said
- Preamble before the actual point
- Bullet points that could be one sentence, or one sentence that could be a clause
Over-qualification
- Hedges that don't add information ("generally", "in most cases", "potentially")
- Disclaimers for obvious or non-applicable scenarios
- Caveats that apply to everything
Scope creep
- Steps or features added "just in case"
- Hypothetical future requirements treated as current
- Handling edge cases that cannot occur in the actual context
Over-engineering (code)
- Helper functions for one-shot operations
- Configurable parameters with only one value ever passed
- Interfaces with one implementation
- Abstractions with only one concrete user
- Error handling for paths the caller guarantees won't happen
Gold-plating (plans/docs)
- Sections that exist to look thorough, not to inform decisions
- Alternatives listed but not needed for the reader to act
- Process steps whose output nobody reads
How to audit
- For each overbaked passage, identify which category applies.
- Propose the tightest rewrite that preserves all load-bearing meaning.
- If something is fine, skip it — do not manufacture findings.
Output format
State the overall verdict in one sentence (lean / slightly overbaked / overbaked / bloated).
Then for each finding:
[Category] — [short label]
original passage or description of the pattern
Rewrite: tighter version
Why cut: one-line explanation of what was lost vs. what was preserved.
Archive
Unless --no-archive was passed: after emitting the audit, write it verbatim to <repo-root>/.ai/reviews/<YYYY-MM-DD>-overbaked[-<target-slug>].md (mkdir -p "$(git rev-parse --show-toplevel)/.ai/reviews"; suffix -2, -3… on filename collision). Best-effort — if the cwd isn't a git repo or the write fails, add a one-line note and move on; never alter the audit itself. Probe the ignore with the file path at repo root, not the bare directory: if git check-ignore -q "$(git rev-parse --show-toplevel)/.ai/reviews/probe.md" exits non-zero, warn and suggest adding .ai/ to .gitignore. (git check-ignore -q .ai false-negatives for the directory-form .ai/ pattern whenever the directory does not yet exist.)
Provenance header. Prepend a YAML block to the archived file only (never to the report shown to the user), so a report re-read months later is still interpretable — in particular, so you can tell whether it predates a revision of this skill's criteria:
---
skill: overbaked
skill-version: <toolkit version>
reviewed-repo: <repo basename> @ <short SHA>[ (dirty)]
cli: <claude --version>
date: <YYYY-MM-DD>
---Collect the values with:
SD="${CLAUDE_PLUGIN_ROOT:+$CLAUDE_PLUGIN_ROOT/skills/overbaked}"; SD="${SD:-$HOME/.claude/skills/overbaked}"
cat "$SD/.version" 2>/dev/null || echo unknown # skill-version ("unknown" under plugin install)
git rev-parse --short HEAD 2>/dev/null || echo unknown
git status --porcelain 2>/dev/null | head -1 # non-empty → append " (dirty)"
claude --version 2>/dev/null || echo unknown.version is written into the deployed skill dir by sync.sh push; it is absent under a
plugin install, which is expected. Resolve any value that fails to unknown — never drop
the key, and never let a failed lookup block or alter the archive.
Before auditing
You need two things: the target material and the intended audience.
- If either is missing, ask for both in one question.
- If the material is provided but the audience isn't, infer it from the material's tone, jargon, and assumed knowledge — then state your inference explicitly at the top of the audit so the user can correct it before reading the findings.
Use the audience to calibrate every finding. What's gold-plating for an expert is essential scaffolding for a novice. State which lens you're applying.
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.

