session-wrap_skill
- Python
460
GitHub Stars
1
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill team-attention/plugins-for-claude-natives --skill session-wrap- SKILL.md6.4 KB
Overview
This skill wraps up a coding session with a structured multi-agent analysis and guided action selection. It inspects git status, summarizes what changed, extracts learnings, finds automation opportunities, validates proposals for duplication, and then asks which actions to perform. The result is a concise set of documentation updates, automation suggestions, learning points, and optional commit or creation actions.
How this skill works
First, it checks git status and recent diffs to understand uncommitted work. Then four analysis agents run in parallel: doc-updater, automation-scout, learning-extractor, and followup-suggester. A validation agent reviews those proposals to detect duplicates or merge needs, after which results are integrated and presented to you. Finally, you choose which actions to execute (commit, update docs, create automation, or skip) and the skill performs only the selected steps.
When to use it
- At the end of a focused coding session to capture what changed and what to do next
- Before switching contexts or handing work off to another developer
- After finishing a feature, bugfix, or refactor to document decisions and learnings
- When you want to discover automation opportunities from repetitive changes
- When you need a validated, low-noise list of follow-up tasks and doc updates
Best practices
- Run the skill with a clear session summary or let it infer changes from git status
- Keep commits small and descriptive; provide a commit message to auto-commit when ready
- Review the doc-updater suggestions before applying to avoid redundant content
- Use the duplicate-checker feedback to merge proposals with existing documentation
- Select only the actions you want executed; the skill will not perform unapproved steps
Example use cases
- Wrap a two-hour development sprint: extract learnings, propose docs updates, and commit changes
- Finish a bugfix and ask for automation ideas to prevent recurrence in future work
- Consolidate decisions and TILs into CLAUDE.md/context.md before ending the day
- Validate proposed follow-up tasks and prioritize what to pick up in the next session
- Quickly decide whether to create a new automation or merge it with an existing one
FAQ
It runs a short status check and a recent-diff summary to identify modified, added, or uncommitted files to include in the session summary.
Can I skip certain agents or phases?
You can control the workflow by providing an explicit commit message to skip interactive prompts, or by choosing which final actions to execute after analysis.