← ClaudeAtlas

retrospective-weeklylisted

Run the weekly devflow self-improvement loop locally: scan freshly-merged watched-author PRs, write per-PR retrospective entries (LLM only for PRs that fail the mechanical clean-gate), derive recurring patterns, and open one human-reviewed intervention PR per actionable pattern. Use when running the weekly devflow retrospective + audit.
The01Geek/devflow-autopilot · ★ 3 · Code & Development · score 69
Install: claude install-skill The01Geek/devflow-autopilot
# /devflow:retrospective-weekly — Weekly Orchestrator This skill is the single entry point the maintainer invokes once a week (or on demand). It is a *conductor*: it runs deterministic bash/jq scripts from `lib/` at every mechanical step and dispatches LLM subagents only at the two genuine-judgment points — per-PR retrospective analysis (Stage A) and per-pattern intervention drafting (Stage B). Everything else — fetching, signal computation, gating, pattern math, and git/PR mechanics — is done by plain scripts with no LLM tokens. ``` LIB="${CLAUDE_SKILL_DIR}/../../lib" ``` All scratch files live under `.devflow/tmp/` (gitignored). Learnings files (`.devflow/learnings/`) are tracked and committed via the state PR. **GitHub autolink hygiene** (any text you compose that lands on a GitHub surface — issue/PR titles, the state-PR report comment, body content you assemble): never put a bare `#` immediately before a number unless it is a real issue or PR reference — GitHub renders `#2` as a link to issue/PR 2, which misleads readers. For an ordinal, count, or list position, spell it out ("item 2", "step 3"), never `#2`. Genuine references like `#123` stay as-is. --- **Consumer prompt extension (load first).** Before doing this skill's work, load any consumer-supplied prompt extension for this skill and honor it. From the repo root, run: ```bash ${CLAUDE_SKILL_DIR}/../../scripts/load-prompt-extension.sh retrospective-weekly ``` If the helper exits non-zero, a consumer extensio