trove-show-work
Keep a reviewable decision trail for long-running or unattended work: an append-only TSV log with one row per decision (what, why, evidence, result). Use for autonomous or multi-phase runs, or work a human reviews after stepping away.
- License 许可
- MIT
- 所属仓库
- bravew/trove
- 最近更新
- 2026-11-12
- Stars
- ★ 10
- Forks
- 1
01 前言 Front Matter
- name
- trove-show-work
- metadata
- { "trove-category": "engineering", "trove-category-label": "Engineering" }
- description
- Keep a reviewable decision trail for long-running or unattended work: an append-only TSV log with one row per decision (what, why, evidence, result). Use for autonomous or multi-phase runs, or work a human reviews after stepping away.
- when_to_use
- keep a decision log; show your work; audit trail for this run
- user-invocable
- true
02 资源文件 3
- 参考文档references/decision-log-template.tsv445 B
- 脚本scripts/log.sh1.1 KB
- 其他SKILL.md.tmpl1.8 KB
03 SKILL.md 正文
<!-- AUTO-GENERATED from SKILL.md.tmpl — do not edit directly --> <!-- Regenerate: bun run build:skills -->
Session Init
This skill ships Trove conventions. Prefer existing project patterns over generic best practices when they conflict.
trove-show-work
Maintain one canonical decision log so a reviewer can reconstruct an unattended run without replaying it. Local by default; commit it when a reviewer needs the trail to trust the result.
The log
One TSV file, append-only, one row per decision. Columns: ts, phase, decision, why, evidence, result. TSV because GitHub renders it as a sortable table and column -s$'\t' -t file.tsv reads it in a terminal. Template: references/decision-log-template.tsv.
Use the helper so timestamps and the formula-injection guard are consistent:
scripts/log.sh decisions.tsv "<phase>" "<decision>" "<why>" "<evidence>" "<result>"
- Append-only. A wrong call doesn't get edited away — add a new row that supersedes it and say so in
result. The history is the point. - Evidence is mandatory. The
evidencecell points at the real thing (a command output, a file:line, a metric), not "looks right".
Self-audit before handoff
Read the log against what actually happened this session. Fix the log, not the story: if a row claims an outcome the transcript doesn't support, correct the row. Optionally get a second opinion from a fresh read (a sub-agent) and add an "Attention" row for anything a reviewer must look at.
Why not prose
A narrative paragraph hides the decision points. One row per decision makes a skipped verification or an unsupported claim visible at a glance.
04 同仓库更多 Skills
using-trove
Skill discipline anchor for the Trove marketplace. Loaded once per session; routes the agent to the right trove-* skill before code, planning, review, or git actions.
trove-write-skill
Skill authoring workflow. Use when creating or updating a Codex or Trove skill, including rubrics, tasks, metadata, and validation.
trove-worktree
Git worktree workflow for spikes, risky experiments, and broad refactors that should not disturb the user's current branch.
trove-why
Reconstruct design rationale and decision history from evidence — source control, issue tracker (Linear), docs (Notion), chat (Slack), error tracking (Sentry) — with calibrated confidence. Use for "why does X work this way", "why did we pick Y", regressions, and postmortems. Use trove-explain for how it works now.
trove-verify
Completion verification workflow. Use before claiming done, fixed, complete, ready, or shipped.
trove-unslop
Cut AI tells from writing — PR descriptions, release notes, specs, commit bodies, and replies. Use before finalizing prose, or when the user asks to clean up, de-slop, or make text sound human.