Borrowing it
Nothing to install: this file belongs to nerds-odd-e/doughnut. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/nerds-odd-e/doughnut/main/.agents/skills/execution-retrospective/SKILL.mdgit clone --depth 1 https://github.com/nerds-odd-e/doughnutWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/nerds-odd-e/doughnut/execution-retrospective)<a href="https://agentmods.dev/skills/nerds-odd-e/doughnut/execution-retrospective"><img src="https://agentmods.dev/badge/skills/nerds-odd-e/doughnut/execution-retrospective.svg" alt="Measured on agentmods" height="20"></a>What 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 5.1 | $0.00089 | $0.02576 |
| Opus 5 | $0.00044 | $0.01288 |
| Sonnet 5 | $0.00018 | $0.00515 |
| Haiku 4.5 | $0.00009 | $0.00258 |
Grade A, and why
execution-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 yesterday.
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.
How it starts
The opening of the file, as written. The whole thing — 264 lines — stays where its author put it; the contents beside it link to each section on GitHub.
When the current conversation contains the execution being reviewed, also review the thread for process improvements and overlooked developer decisions. Do not implement findings, edit rules or skills, commit, push, or start plan execution.
The PLAN may have been removed by normal completed-plan cleanup. A partial capability name, surviving conversation context, old path, commit message, or distinct story phrase is sufficient input for discovery. Do not require the user to supply a deleted file's exact name.
The review is read-only except for amending the unfinished PLAN or creating a follow-up PLAN through slice-planning. Preserve existing working-tree changes when making those planning edits. Do not create a separate retrospective artifact unless the user asks.
- Current conversation and execution transcript.
- Current
.planning/phases/and.planning/quick/contents. - Git history for renamed or deleted planning paths.
- Commit messages and diffs containing distinctive story language.
Useful history operations include:
git log --all --name-status -- .planning/phases .planning/quick
git log --all --diff-filter=D --summary -- .planning/phases .planning/quick
git log --all --grep='<capability phrase>'
git log --all -S'<distinct story phrase>' -- .planning
git show <revision>:<historical-plan-path>
Recover the earliest execution-ready plan revision, its source seed/story when present, beneficiary, intended outcome, boundaries, and outside-in proof. Also record later plan changes that were explicitly approved or supported by new evidence; do not misclassify them as drift.
Determine completion from the latest PLAN's slice statuses and execution
evidence, not the earliest revision or the file's presence alone. Any planned
or in-progress slice means the plan is unfinished. Record completed slices and
remaining work separately; a deleted PLAN requires history evidence of
completion.
If two plans remain equally plausible after history inspection, stop and ask the developer to choose. Do not combine them.
- it updates the PLAN's slice status or cleanup;
- its message names the story, capability, or slice;
- its diff implements, proves, refactors, or documents the recovered story; or
- the execution thread explicitly associates it with the plan.
Record each included SHA and reason. Inspect intervening commits and exclude unrelated work; do not assume one contiguous range belongs to the plan.
Review all selected commit diffs together. When they are contiguous and contain no unrelated work, use the net diff from the parent of the first implementation commit through the last related commit. Otherwise review the selected patches together and inspect their changed files at the last related implementation commit. Do not let later commits or current uncommitted work contaminate the historical result, and never mutate the current worktree to reconstruct it. Keep planning-only changes as provenance, not as product-quality findings.
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.
- yesterday First seen · 264 lines · 89 tokens per session scan A 4d0c9600f590
execution-retrospective is a skill published in the GitHub repository nerds-odd-e/doughnut (49 stars, last pushed today), licensed MIT. It adds 89 tokens to every session and 2,576 once invoked, about $0.0004 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 2026-09-05.
Other skills, from other repositories
autoreview
Pre-commit/ship code review: Codex default; optional Claude or Pi.
rework-rate
Measure and interpret PR rework rate — the emerging 5th DORA metric.
omh-code-review
This is a Hermes-native code-review workflow skill.
revdiff-plan
Review the last Codex assistant message (plan, analysis, or proposal) with inline annotations in a TUI overlay. Extracts the most recent response from Codex rollout files and opens it in revdiff for review and annotation. Activates on "revdiff-plan", "review plan with revdiff", "annotate plan", "review last response"…
code-reviewer
Code review specialist focused on patterns, bugs, security, and performance.
full-repo-review
Comprehensive four-wave review of all repo source files, producing a prioritized issue backlog.