
Commit History
List recent git commits tied to agent sessions with SHA, branch, message, and observation counts via AgentMemory MCP or HTTP.
Overview
commit-history is an agent skill most often used in Operate (also Ship, Grow) that lists git commits linked to agent sessions via AgentMemory with optional branch and repo filters.
Install
npx skills add https://github.com/rohitg00/agentmemory --skill commit-historyWhat is this skill?
- Parses branch=, repo=, and limit= from arguments (default limit 100, max 500)
- Primary path: memory_commits MCP tool; HTTP fallback to GET /agentmemory/commits with URL-encoded filters
- Bearer AGENTMEMORY_SECRET when set; never invents commits on empty results
- Renders short SHA, branch, timestamp, message line, session ids (8 chars), observation and file counts
- Default commit list limit 100 with max 500
- Session ids displayed as first 8 characters when present
Adoption & trust: 1.2k installs on skills.sh; 21.8k GitHub stars; 2/3 security scanners passed (skills.sh audits).
What problem does it solve?
You cannot tell which commits came from which agent run or how much context was attached when the user asks what the agent shipped.
Who is it for?
Builders using rohitg00/agentmemory who want session-attributed shipping history from chat (“show agent commits”).
Skip if: Repos without AgentMemory linked sessions or users who only need raw git log with no agent correlation.
When should I use this skill?
User asks to show agent commits, what the agent shipped, or wants commit lists with session context; argument-hint supports branch, repo, limit.
What do I get? / Deliverables
You receive a reverse-chronological, filterable commit list with session ids and observation metadata—and explicit empty-state guidance instead of fabricated history.
- Formatted reverse-chronological commit list with session and observation metadata
- Empty-result message with filter relaxation suggestions
Recommended Skills
Journey fit
Spans multiple journey phases - primary shelf plus alternate fits below.
Operating agent-assisted repos requires visibility into what the agent actually shipped versus human commits. Monitoring subphase fits session-linked commit feeds—reverse-chronological audit of agent output with filters.
Where it fits
Filter commits on main after an overnight agent run to see SHAs and linked session ids.
Before tagging a release, confirm only expected agent sessions produced commits in the repo fragment.
Pull limit=500 history to count how many agent-attributed commits landed this week per branch.
How it compares
Session-aware commit index via MCP/HTTP—not a replacement for git blame or conventional CI deployment dashboards.
Common Questions / FAQ
Who is commit-history for?
Solo developers and agent operators on AgentMemory who audit agent-driven git activity by branch or repository.
When should I use commit-history?
Use in Operate when reviewing what agents changed in production repos, in Ship after a release push to verify session-linked commits, and in Grow when summarizing shipping velocity for yourself or stakeholders.
Is commit-history safe to install?
It reads commit metadata through your AgentMemory endpoint and may use AGENTMEMORY_SECRET; review the Security Audits panel on this Prism page and scope tokens to least privilege.
SKILL.md
READMESKILL.md - Commit History
The user wants a list of agent-linked commits. Filter args: $ARGUMENTS Parse `$ARGUMENTS` for optional `branch=<name>`, `repo=<url-or-fragment>`, and `limit=<n>` tokens. A bare numeric token becomes the limit. Defaults: no branch filter, no repo filter, limit 100, max 500. Call the `memory_commits` MCP tool with the parsed filters. If the MCP tool is unavailable, fall back to HTTP: build `GET $AGENTMEMORY_URL/agentmemory/commits` and append each filter as a URL-encoded query parameter (use `URLSearchParams` or `encodeURIComponent` on `branch`, `repo`, and `limit`) so values containing `?`, `&`, or `#` cannot corrupt the request. Include `Authorization: Bearer $AGENTMEMORY_SECRET` when set. Render the result as a reverse-chronological list: - Short SHA, branch, authored timestamp - Commit message first line - Linked session id(s) (first 8 chars each) and observation counts where present - File count when `files` is provided If the result is empty, tell the user the filter matched no commits and suggest dropping the branch/repo filter. Do not invent commits.