recap
Skill3 starsChanged 42 days agoMITView on GitHub
Plain-language summary of what changed and why — for a branch, PR, uncommitted work, or a commit range. Explains the causal chain from the original goal to where things stand now, grouped by intent rather than by file, plus status and next steps. Use for "recap", "what have we changed so far", "summarize this PR vs main", "catch me up on this branch", or a wrap-up at the end of a feature.
44 other projects use a version of this.
Adopted by
Most recently taken up by
- mel0nyrame/skillstheir own version, first seen Nov 8, 2026
- delorenj/skillstheir own version, first seen Nov 7, 2026
- Integralist/agent-skillstheir own version, first seen Nov 6, 2026
- acegalaxy-co/ace_commons-ccswitch-clitheir own version, first seen Nov 1, 2026
- kanalias/ace_commons-ccswitch-clitheir own version, first seen Nov 1, 2026
- AcKeskin/contexturetheir own version, first seen Nov 1, 2026
Used with
Projects with it also use glossary, orchestrate and 7 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.
Microsoft AzureSkillglossary
Short glossary of domain terms used by the team.
Also in 3 projects that use this
varun29ankuSSkillorchestrate
This skill should be used when the user asks to 'orchestrate a task', 'break down work into parallel agents', 'coordinate subtasks', 'run agents in parallel', or mentions 'multi-agent'. Decomposes complex tasks into tracked subtasks, dispatches parallel subagents, and coordinates until completion.
Also in 4 projects that use this
sunnyclinches44Skillexplain
Explains code using real-world analogies and ASCII diagrams
Also in 3 projects that use this
Microsoft AzureSkillbuild
Build the TypeSpec Rust emitter from source. Run this after cloning the repo, after pulling changes, or when TypeScript source files have been modified.
Also in 3 projects that use this
nicolasestremSkillbrainstorm
Structured idea generation and solution exploration with systematic analysis.
Also in 3 projects that use this
BlockSkillplan
Draft a TDD, PRD, or lightweight plan in the conversation without committing it.
Also in 3 projects that use this
Other versions44 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 44
- first-fluke/fullstack-starter★ 226same name
- backnotprop/bro★ 176same name
- benoror/obsidianos_work★ 166same name
- Soul-Brews-Studio/arra-oracle-skills-cli★ 121same name
- mbwsims/claude-universe★ 71same name
- TakaGoto/rag-learning-academy★ 24same name
- Integralist/agent-skills★ 23same name
- delorenj/skills★ 17same name
- ivklgn/ai-kit★ 12same name
- ivklgn/bag-of-skills★ 12same name
34 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 repository4 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- public/llms.txtllms.txt
- panel-reviewSkill
The file
We split this skill into the rules it gives an agent
3 rules. 0 are common, 3 are distinctive, and 0 are very specific to this project.
0
Common
Most of the 38 edited versions we read keep these.
3
Distinctive
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 17
Run this inline in the current session.
- Distinctive: line 96
Plain words: Avoid function, table, and API names unless the name *is* the point.
- Distinctive: line 106
Flag the unverified: If something is untested, assumed, or known- incomplete, say so in Status.
“Common” is measured: kept by at least 6 in 10 of the 38 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
Recap
Explain a body of work to the person who owns it, in the terms they'd use. Two moments this is for: mid-development, when the thread of what's been done and why has gotten long, and end of feature, when reviewing the whole change before merging.
This is not a changelog. git log already exists and is useless for this
purpose. The job is to reconstruct intent — what was wanted, what that
turned up, what it turned into — and state where things stand.
Run this inline in the current session. No subagents: there's no judgment to keep independent here, and if this session did the work, its memory of the reasoning is the single best source for the "why".
Scope
- (bare) — current branch vs its merge-base with the default branch.
<PR#>— that PR vs its base (gh pr view <n>,gh pr diff <n>).staged/working— uncommitted changes only.<ref>..<ref>— an explicit commit range.since <date/ref>— work since a point in time.
Quote paths in shell commands — parens and spaces (e.g. route groups like
app/(main)/...) break unquoted.
Gather before writing
Cheap, in rough priority order. Stop when you can explain the why, not just the what:
- This conversation, if the work happened here — the strongest source for intent and for the causal chain between changes.
- Commits: messages and order tell the story of how the work evolved
(
git log --oneline,git diff --stat). - The PR description, if one exists.
- Artifacts that record intent:
reviews/*-review.md(a panel review's finding counts, dispositions and final verdict),plans/,TODO.mdentries, and anydocs/file the diff itself modified — docs changed inside a diff usually state the design intent outright. - The code, last — to confirm what the above claims, and to catch anything undocumented.
If the "why" genuinely can't be reconstructed, say so ("the commits don't record why X was changed") rather than inventing a plausible rationale. A recap that quietly guesses at intent is worse than one that admits a gap.
Output shape
Print it in chat. Don't write a file unless asked.
<One line: scope vs base, and the scale — "Five commits, ~8 files.">
## What it does
1. **<Headline: the effect, in plain words>**
<2–4 sentences. Lead with the symptom or capability a person would
notice. Then just enough mechanism to make it make sense. Anchor to a
concrete event where one exists.>
2. **<Next one>**
<...>
## Why
<The causal chain in 2–4 sentences: what was originally wanted → what that
surfaced → what it became. This is the section people actually re-read.>
## Status
<Where it stands, what's left, and what to do next — in order, including
any deploy/migration sequencing the changes imply.>Group items by intent — a bug fix, a new capability, fixes from review — not by file or by commit. Three to five items is usually right; if you have ten, you're listing changes instead of explaining work.
Writing rules
These are what separate a good recap from a diff summary:
- Effect first, mechanism second. "New skill content showed up blank" before "the cache tag was pinged before content landed". Someone should understand the problem before they meet the machinery.
- Anchor to real events. "This is exactly what happened with improve-ui" turns an abstract bug into one the reader remembers. Use these whenever the history supplies one.
- Plain words. Avoid function, table, and API names unless the name is the point; if a term is unavoidable, gloss it in the same sentence. A reader who's been away for a week should follow it without opening the code.
- Numbers beat adjectives. "2 rounds, 14 findings, all resolved" says more than "extensive review".
- Under a minute to read. Roughly 400 words. Cut ruthlessly — length is the most common way this goes wrong.
- Report, don't sell. No "great work", no "robust solution". State what is.
- Flag the unverified. If something is untested, assumed, or known- incomplete, say so in Status. This is often the most valuable line.
Anti-patterns
- A file-by-file walkthrough (that's
git diff --stat). - Restating commit messages in order.
- Code blocks or line numbers — this is prose; cite a path only when the reader would need to go there.
- Explaining what the code does while never saying why it exists.
- Padding a small change into a long summary to look thorough.
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.

