session-harvestlisted

Use when invoked explicitly as /session-harvest, or when the user asks what's worth saving before compacting/ending a session, or says something like 'harvest this session', 'anything to remember here', 'anything dangling before I stop', or 'is it safe to compact'. Reviews the conversation for anything worth keeping and routes each item to a plain file every agent can read: plan-specific content to plans/*.md (per the plan-docs skill), repo-specific durable knowledge to that repo's AGENTS.md/docs/contributing, and cross-repo/personal preference to ~/AGENTS.md. Never a harness's own memory store, for any project or any reason — that vendor-locks the work. Then sweeps live state the conversation can't show: processes the session left running, unpushed commits in every repo it touched, CI on what it pushed, and work it promised but never verified. Ends with a report ordered least- to most-urgent, a safe-to-compact verdict, and a next-session prompt. On-demand only — never installs hooks or runs automatically.
TheodoreAD/agent-skills · ★ 0 · AI & Automation · score 75

Install: claude install-skill TheodoreAD/agent-skills

# Session harvest Reviews a conversation for what's worth persisting before it's compacted or ends, and makes sure each candidate lands in the _right_ place — not just memory by default because memory is the easiest tool at hand. Primary invocation is explicit (`/session-harvest`); the description above also matches natural-language phrasing, but explicit invocation is the one to rely on — description matching for something this consequential (it writes files) has been reported unreliable by prior art (see `references/rationale.md`). On-demand only. Installing hooks, wiring `settings.json`, or running on any schedule is explicitly out of scope — that's a different, heavier tool (see `references/rationale.md` for the prior art considered and rejected). ## What this skill reads, runs and writes - **Reads**: Claude Code's transcripts and background-job state under `~/.claude` (read-only, another harness has nothing there); git state of every repo the session touched; CI runs through `gh`; the process and listening-socket tables; docker images; both plans stores; the installed skills hub and the source checkout `skills-state` resolves. The socket step checks whether credential-named files (`.env`, `.netrc`, …) **exist** under a directory a listener serves — names only, never contents. - **Runs**: `git` (including `fetch` to each touched repo's own upstream unless `--no-fetch`), `gh` and `docker` when installed, `ps`/`ss` on Linux or PowerShell/`netstat` on Windows