task-feedback-analyserlisted

# Task Feedback Analyser Extract recurring patterns from completed task feedback into generalizable workflow rules. > Part of the task planning workflow. See [references/workflow-overview.md](references/workflow-overview.md) for the full pipeline. ## Process ### Step 1: Scan feedback sources - Read all `docs/tasks/*/feedback.md` files - Read review comment threads from completed PRs (merged) - Track which files have been processed to avoid re-scanning (check against existing rules file for "Evidence" task references) ### Step 2: Cluster by theme Group observations by theme. Examples of themes: - Missing tenant isolation - Stale context references - Tests that don't verify actual behavior - Incomplete error handling in server functions ### Step 3: Filter to 3+ independent examples A theme qualifies as a rule only when it appears in **3 or more independent examples** (different task folders). This minimum is strict -- never fabricate or stretch examples to meet the threshold. Drop themes that don't meet the bar. Report dropped themes so the user knows they exist but aren't mature enough. ### Step 4: Draft rules Write each rule in imperative form with a `use_when` trigger and verbatim evidence: See [references/template-rule.md](references/template-rule.md) for the rule template format. ### Step 5: Update rules file - Read existing `docs/rules.md` (or create if it doesn't exist) - Merge new rules with existing ones -- do not duplicate - Assign sequential IDs conti