wrap-up
Skill65 starsChanged 7 months agoMITView on GitHub
End-of-conversation shipping and review workflow. Use when the user says wrap up, close out, end session, ship this, commit and push, prepare handoff, or asks for a session closeout. Runs a scheduler-safe preflight, verifies work, commits and pushes specific files when appropriate, captures open loops and improvement candidates, and ends with a clear handoff.
64 other projects use a version of this.
Adopted by
Most recently taken up by
- flegaspi700/pmo-project-standardizationtheir own version, first seen Nov 16, 2026
- manavpurswani/claude-wrap-uptheir own version, first seen Nov 16, 2026
- mm-arturo/claude-code-skillstheir own version, first seen Nov 15, 2026
- PaulRBerg/agent-skillstheir own version, first seen Nov 13, 2026
- bas1l/claude-skillstheir own version, first seen Nov 11, 2026
- adeonir/agent-skillstheir own version, first seen Nov 10, 2026
Used with
Projects with it also use new-project, refine and 9 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.
zcagSkillnew-project
Guide users through creating a new software project from scratch — from raw idea to actionable tasks. Use this skill whenever someone says "new project", "start a project", "create a project", "scaffold a project", "I want to build X", "help me plan a project", "I have an idea for an app/tool/service", or mentions they're starting something new. Also trigger when someone asks for help with project planning, tech stack decisions, MVP scoping, feature planning, or figuring out what to build and how to break it down. When the user describes any project idea and wants to move forward on it, this skill should activate.
Also in 3 projects that use this
AlexPEClubSkillrefine
Always use when the user wants to discuss an exsiting feature or specification. Open an existing feature spec to improve, extend, or fundamentally challenge it. Pass the feature ID as argument (e.g. /refine PROJ-2).
Also in 3 projects that use this
MicrosoftSkilldoctor
Diagnose an Amplifier CLI environment through the CLI doctor agent. USE WHEN a user requests CLI runtime, settings or missing-capability inspection. DO NOT USE for project bugs or other hosts.
Also in 4 projects that use this
austintgriffithSkillorchestration
Use when a Scaffold-ETH 2 build leaves localhost: planning the launch order, deploying and verifying contracts on a live network, standing up a fork of a real chain to rehearse a deploy, or fixing a contract bug that is already onchain. Not for running the app or a demo against that fork (`/frontend-playbook`).
Also in 3 projects that use this
MicrosoftSkillsync
Sync the current session branch with its upstream branch, or publish the current session branch to a remote. Use when the user asks to sync a branch, pull latest changes, rebase onto upstream, push current branch, publish branch, or set upstream.
Also in 3 projects that use this
nicolasestremSkillbrainstorm
Structured idea generation and solution exploration with systematic analysis.
Also in 6 projects that use this
- +1 more from microsoft
Other versions64 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 64
- rohitg00/pro-workflow★ 2.9ksame name
- anthropics/launch-your-agent★ 1ksame name
- codewithmukesh/dotnet-claude-kit★ 808same name
- ianlkl11234s/mini-taiwan-pulse★ 547same name
- aspi6246/Claude-Code-Skills-for-Academics★ 153same name
- PaulRBerg/agent-skills★ 104same name
- KerberosClaw/kc_ai_skills★ 79same name
- simoncorry/foundry★ 39same name
- alxyrgin/health-os★ 38same name
- mshadmanrahman/claude-for-designers★ 32same name
54 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 repository19 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- agent-browserSkill
- buildSkill
- debugSkill
- financial-advisorSkill
- generate-image-openaiSkill
- good-morningSkill
- pdfSkill
- planeSkill
- reply-xSkill
- researchSkill
- search-historySkill
- send-emailSkill
- simplify-julesSkill
- skill-creatorSkill
- stop-slopSkill
- thinkSkill
- watch-contactsSkill
- writeSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
32 rules. 0 are common, 27 are distinctive, and 5 are very specific to this project.
0
Common
Most of the 59 edited versions we read keep these.
27
Distinctive
This author's own choices. Worth a look.
5
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- could be automatic: a setting could enforce it
- Distinctive: line 8
Use this at the end of a meaningful conversation or when the user asks to ship, close out, commit, push, or prepare a handoff.
- Very specific: line 17
Start with Workflow: wrap-up.
- Very specific: line 24
If this skill is running from a copied surface where ~/.agents is unavailable, run the copy-local script at scripts/wrapup_preflight.py.
- Distinctive: line 29
Run these phases in order.
- Distinctive: line 40
Do not call the session complete if the stated objective is unfinished.
- Distinctive: line 44
Use the preflight script output as the collision map.
- Distinctive: line 55
Do not restart, unload, kickstart, repair, or create background services during normal wrap-up.
- Very specific: line 57
Do not edit LaunchAgents, Hermes config.yaml, Hermes cron/jobs.json, secrets runtime files, scheduler scripts, or automation definitions …
- Distinctive: line 60
Treat scheduler output, logs, state snapshots, lock files, and generated reports as owned by running processes unless they are part of this …
- Distinctive: line 63
If a scheduled job is unhealthy, record it as an improvement candidate unless the user explicitly asked for scheduler repair.
- Distinctive: line 65
If a job is due soon, avoid broad staging.
- Distinctive: line 68
If the preflight cannot inspect a surface because of permissions, report that surface as unverified.
- Distinctive: line 73
Run the narrowest meaningful checks for the work done:
- Distinctive: line 81
If verification is skipped, state why.
- Distinctive: line 100
Never use git add . or git add -A.
- Distinctive: line 104
Commit only after verification and change review.
- Distinctive: line 106
Use a descriptive message that says what shipped, not that the session ended.
- Distinctive: line 115
Avoid vague messages like:
- Very specific: line 123
Do not use --no-verify unless the user explicitly approves bypassing hooks and the reason is stated in the final report.
- Distinctive: line 133
Do not force push.
- Distinctive: line 134
If push is rejected, report the reason and stop for review unless the user already authorized pull/rebase behavior.
- Distinctive: line 136
Do not pull or rebase during generic wrap-up unless that behavior was clearly approved for this session.
- Distinctive: line 138
If the user said "do not pull", preserve that constraint.
- Distinctive: line 156
Look for improvements in these categories:
- Distinctive: line 164
Do not automatically edit shared instructions during wrap-up.
- Distinctive: line 200
If the objective continues, end with:
- Distinctive: line 224
Do not save a file for trivial Q&A, quick lookups, or conversations with no state change.
- Distinctive: line 227
When saving, use a run artifact path rather than a durable Knowledge-OS report:
- Distinctive: line 234
Use the helper to create the artifact deterministically:
- Distinctive: line 268
Use the read-only target planner before installing or changing surface links:
- Very specific: line 283
Do not install this skill on archivist, dreamer, researcher, or subconscious unless Jonathan explicitly re-approves those profile surfaces.
- Distinctive: line 289
Do not replace existing physical skill directories with symlinks without review.
“Common” is measured: kept by at least 6 in 10 of the 59 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
Wrap-Up
Use this at the end of a meaningful conversation or when the user asks to ship, close out, commit, push, or prepare a handoff.
The job is not just repo cleanup. The job is to leave the workspace in a known state, avoid stepping on running schedulers, ship only intentional changes, and turn repeated friction into improvement candidates.
Opening
Start with Workflow: wrap-up. Then run the deterministic preflight before
making any changes:
python3 ~/.agents/skills/workflows/wrap-up/scripts/wrapup_preflight.py --workspace "$PWD" --due-window-minutes 90If this skill is running from a copied surface where ~/.agents is unavailable,
run the copy-local script at scripts/wrapup_preflight.py.
Core Sequence
Run these phases in order.
1. Outcome State
Summarize the session's original objective and current state:
- what got done
- what did not get done
- whether the work changed scope
- whether the handoff is
CompleteorContinues
Do not call the session complete if the stated objective is unfinished.
2. Scheduler-Safe Preflight
Use the preflight script output as the collision map. At minimum, account for:
- user LaunchAgents in
~/Library/LaunchAgents - user crontab from
crontab -l - Hermes cron jobs in
~/.hermes/profiles/*/cron/jobs.json - Codex automations in
~/.codex/automations/*/automation.toml - running Hermes gateways, dashboards, WebUI, Dev Secrets, history-search, and other launchd-managed services when visible
Default safety rules:
- Do not restart, unload, kickstart, repair, or create background services during normal wrap-up.
- Do not edit LaunchAgents, Hermes
config.yaml, Hermescron/jobs.json, secrets runtime files, scheduler scripts, or automation definitions unless the session explicitly touched scheduler work. - Treat scheduler output, logs, state snapshots, lock files, and generated reports as owned by running processes unless they are part of this session's deliverable.
- If a scheduled job is unhealthy, record it as an improvement candidate unless the user explicitly asked for scheduler repair.
- If a job is due soon, avoid broad staging. Stage only files known to belong to this session.
If the preflight cannot inspect a surface because of permissions, report that surface as unverified. Do not claim it is clean.
3. Verification
Run the narrowest meaningful checks for the work done:
- tests for code changes
- build or typecheck for frontend/app changes
- script smoke checks for wrappers and automation helpers
- syntax checks for JSON, TOML, YAML, plist, shell, or Python files
- health checks for services only when the session touched those services
If verification is skipped, state why. A skipped check is not a pass.
4. Change Review
Inspect git before staging:
git status --short
git diff --statClassify changes:
- session changes to stage
- user or scheduler changes to leave alone
- generated artifacts to review before staging
- secret or credential-looking files to avoid and flag
- unrelated dirty files to preserve
Never use git add . or git add -A. Stage specific files only.
5. Commit
Commit only after verification and change review.
Use a descriptive message that says what shipped, not that the session ended. Prefer:
Add scheduler-safe wrap-up workflow
Repair Pam calendar auth probe
Update Jules Live Studio turn handlingAvoid vague messages like:
Wrap-up changes
Session updates
Misc fixesDo not use --no-verify unless the user explicitly approves bypassing hooks and
the reason is stated in the final report.
6. Push
Push the current branch to its configured upstream or to the branch requested by the user.
Rules:
- Do not force push.
- If push is rejected, report the reason and stop for review unless the user already authorized pull/rebase behavior.
- Do not pull or rebase during generic wrap-up unless that behavior was clearly approved for this session.
- If the user said "do not pull", preserve that constraint.
7. Open Loops and Commitments
List only real open loops:
- blockers
- next actions
- user commitments to another person
- outbound drafts needing approval
- calendar, email, social, or partner follow-ups
- background jobs or health issues discovered but not repaired
For a continuing handoff, include a resume prompt specific enough for a fresh session.
8. Improvement Candidates
Look for improvements in these categories:
- Skill gap: the assistant struggled or missed existing guidance.
- Friction: a repeated manual step should become a script, check, or skill.
- Knowledge: a durable fact should be captured in the right local doc.
- Automation: a scheduled job, wrapper, or health check could be hardened.
- Quality: tests, docs, or validation were missing for behavior that matters.
Do not automatically edit shared instructions during wrap-up. Present concrete candidates with suggested files or procedures.
Hermes profiles may run their own profile-native skill review/update procedure.
Codex and Claude should follow this skill's guidance and propose shared skill
edits for review before changing .agents or profile-specific skills.
9. Final Report
End with a compact report:
Wrap-up complete.
Verification:
- [check]: pass/fail/skipped
Git:
- branch:
- commit:
- push:
Scheduler safety:
- checked:
- due soon:
- unverified:
Open loops:
- ...
Improvement candidates:
- ...
Handoff: Complete - [one sentence]If the objective continues, end with:
Handoff: Continues - [what remains]
Resume prompt: [exact prompt]On Codex, after a successful stage, commit, push, branch creation, or PR creation, emit the relevant Codex app git directives in the final answer.
10. Optional Report Artifact
The conversational final report is always required. A file artifact is conditional.
Save a wrap-up report file when at least one is true:
- code, docs, config, or assets were committed or pushed
- a durable decision was made
- the work continues and needs a fresh-session resume prompt
- a background job, scheduler, automation, auth surface, or service was touched
- there are concrete improvement candidates worth reviewing later
- Jonathan explicitly asks for a report artifact
Do not save a file for trivial Q&A, quick lookups, or conversations with no state change.
When saving, use a run artifact path rather than a durable Knowledge-OS report:
System/Runs/<surface>/wrap-up/YYYY-MM-DD/HHMMSS-<slug>.md
System/Runs/<surface>/wrap-up/YYYY-MM-DD/HHMMSS-<slug>.jsonUse the helper to create the artifact deterministically:
python3 ~/.agents/skills/workflows/wrap-up/scripts/wrapup_report.py \
--workspace "$PWD" \
--surface codex \
--status Complete \
--focus "Add scheduler-safe wrap-up workflow" \
--summary "Created the shared wrap-up skill and helper scripts." \
--verification "preflight helper runs" \
--improvement "Review surface symlink plan before installing adapters" \
--writeReport artifacts are receipts and ingestion inputs. They do not replace transcripts, git history, history-search, Pam reports, Subconscious notes, Researcher intake, Archivist distillation, Builder's Log, or future gbrain.
For gbrain, keep the artifact short and structured: session id if known, surface, status, focus, files/commits, verification, scheduler safety, open loops, improvement candidates, and resume prompt. Do not paste raw transcript content or secret values.
Promote only durable decisions, reusable facts, or stable operating rules into Knowledge-OS or memory. Do not treat every wrap-up artifact as durable knowledge.
Surface Adapters
The canonical skill lives at:
~/.agents/skills/workflows/wrap-upUse the read-only target planner before installing or changing surface links:
python3 ~/.agents/skills/workflows/wrap-up/scripts/wrapup_surface_targets.pyRecommended adapter behavior:
- Codex: symlink
~/.codex/skills/wrap-upto the canonical skill. Codex adds app git directives after successful git operations. - Claude Code: symlink
~/.claude/skills/wrap-upto the canonical skill. Claude uses plain final text and its own slash-command behavior. - Hermes profiles: symlink
~/.hermes/profiles/<profile>/skills/workflows/wrap-upfor approved workflow-owning profiles only. Current live adapters arepamanddirector. Do not install this skill onarchivist,dreamer,researcher, orsubconsciousunless Jonathan explicitly re-approves those profile surfaces. Hermes may run profile-native skill review before proposing shared edits. - Public/reference Jules repo: copy or sync a sanitized copy into the repo rather than symlinking to a host-local path.
Do not replace existing physical skill directories with symlinks without review.
Deterministic Helpers
scripts/wrapup_preflight.py: read-only scheduler, git, and automation preflight. It intentionally avoids prompts and secret values.scripts/wrapup_surface_targets.py: read-only inventory of surface adapter targets and whether each is missing, a symlink, or a physical directory.scripts/wrapup_report.py: optional writer for compact Markdown plus JSON wrap-up artifacts underSystem/Runs/<surface>/wrap-up/.
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.

