tinyhat:history — browse all skill-audit reports

tinyhat:history — browse all skill-audit reports

Open ${CLAUDE_PLUGIN_DATA}/archive/index.html. That page lists the latest report plus every dated archive snapshot (up to 31) with one-click links. Each entry opens its own report.html.

If the new plugin-data path is empty but the legacy path ~/.claude/tinyhat/archive/index.html exists, use the legacy path for this browse. The next write-capable Tinyhat command will migrate it.

Skill-relative script path

${CLAUDE_SKILL_DIR} is rendered into the skill content before Claude runs the Bash block below, so the command keeps pointing at the version of Tinyhat that loaded this skill.

Flow

1. Check whether ${CLAUDE_PLUGIN_DATA}/archive/index.html exists.
   • If it doesn't, check ~/.claude/tinyhat/archive/index.html.
   • If the new path is missing but ${CLAUDE_PLUGIN_DATA}/latest/report.html exists, regenerate the index without running a new audit:

     python3 "${CLAUDE_SKILL_DIR}/../../scripts/render_report.py" --index-only
     

   • If no reports exist at all, tell the user and hand off to /tinyhat:audit to create the first snapshot.
2. Open the index:

open "${CLAUDE_PLUGIN_DATA}/archive/index.html"        # macOS
xdg-open "${CLAUDE_PLUGIN_DATA}/archive/index.html"    # Linux
start "${CLAUDE_PLUGIN_DATA}/archive/index.html"       # Windows

3. In one sentence, note how many dated snapshots are listed.

Navigation inside the browser

Every report page has an all reports → link in its header that returns the user here. From this page, each entry opens its own report.html. No server — pure static files.

When the user wants a fresh report

Direct them to /tinyhat:audit. This skill only browses; it doesn't regenerate.