Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/nerds-odd-e/doughnut/codebase-retrospectivenpx skills add nerds-odd-e/doughnut --skill codebase-retrospectivegit clone --depth 1 https://github.com/nerds-odd-e/doughnutWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00052 | $0.00926 |
| Opus 5 | $0.00026 | $0.00463 |
| Sonnet 5 | $0.00010 | $0.00185 |
| Haiku 4.5 | $0.00005 | $0.00093 |
Grade A, and why
codebase-retrospective scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
Purpose: discoverability and code quality, not task planning (use slice-planning when the task itself is too large).
Output: Short friction report ending with ## RETROSPECTIVE COMPLETE, then
resume the original work.
- You notice repeated confusion, wrong turns, or wasted searches in your session.
- The developer asks for a retrospective on friction or discoverability.
This is about codebase quality, not task planning. If the task itself is too big, use slice-planning.
Take 1–2 minutes — do not derail the session.
- Searched for something and couldn't find it?
- Misunderstood code because of naming?
- Had to read many files to understand one concept?
- Waited a long time for a test or build?
- Couldn't tell which file/class was responsible for a behavior?
| Category | Examples |
|---|---|
| Organization / naming | Misleading file name, concept split across distant files, unclear module boundaries |
| Missing abstraction | Duplicated logic, no shared helper, leaky internal details |
| Missing / stale documentation | No doc for non-obvious convention, outdated comments, missing CLAUDE.md guidance |
| Slow feedback | Slow test suite, long build, manual steps that could be scripted |
- Self-explaining code — Rename, extract well-named helper, move related code together. Almost always better than documentation.
- Automated protection — Test, lint rule, script, or CI check that catches the problem mechanically.
- Documentation — Only when code structure alone cannot convey the information (env setup, cross-repo conventions, architectural rationale). Keep minimal and close to where needed.
Never add redundant comments or docs that repeat what clear code already says.
For each friction point:
- Fix directly if: small improvement (rename, move, extract helper, targeted test, script alias) and won't break existing tests or workflows. Do it now.
- Propose to the developer if: structural (extract module, reorganize directory, new abstraction, speed up test suite). Write a short description in your response.
Be specific. "The code was confusing" is not useful. "I searched for reading progress
logic in 4 files before finding BookReadingContent.vue because the name doesn't
suggest progress tracking" is useful.
Only propose improvements that would have actually saved you time in this session.
- What friction you hit (concrete examples from your session).
- What you fixed directly (if anything).
- What you propose they consider (if anything structural).
Then continue working on the original task.
<success_criteria>
- Specific friction points identified (not vague complaints)
- Each point categorized and addressed (fix or propose)
- Session continues on original task after report
- Final output includes
## RETROSPECTIVE COMPLETE</success_criteria>
- Friction points (concrete examples).
- Direct fixes applied (if any).
- Structural proposals (if any).
## RETROSPECTIVE COMPLETE
Then resume the original task.
<out_of_scope>
- Do not replan the task (use slice-planning for that).
- Do not spend more than 1–2 minutes on the retrospective.
- Do not add documentation that merely repeats clear code. </out_of_scope>
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 121 lines · 52 tokens per session scan A 7887d114fdfe
codebase-retrospective is a skill published in the GitHub repository nerds-odd-e/doughnut (49 stars, last pushed 2d ago), licensed MIT. It adds 52 tokens to every session and 926 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
brainstorming
You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.
chat-pet-sprite-creation
Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…
agent-host-chat-contributions
Build and review cross-cutting agent-host chat behavior through lifecycle contributions. Use when adding turn lifecycle side effects, prompt or context injection, restored-history transformation, protocol-action observation, or when reviewing changes that add code to AgentSideEffects or AgentService.
auto-perf-optimize
Run agent-driven VS Code performance or memory investigations. Use when asked to launch Code OSS, automate a VS Code scenario, run the Chat memory smoke runner, capture renderer heap snapshots, take workflow screenshots, compare run summaries, or drive a repeatable scenario before heap-snapshot analysis.