jira-refactor-findings

# Jira refactor findings Every Jira action for refactor findings goes through these scripts. Don't hand-roll Jira calls. ```bash J=.agents/skills/jira-refactor-findings/scripts pnpm exec tsx $J/<script>.ts [...] ``` Settings live in `config.json`: project, issue type, label `refactor-findings`, the Pod field and its option id, cooldowns, the per-run ticket cap, and the env var holding the token. Exit code 2 means Jira could not be reached or rejected the credentials. Never treat it as a pass. ## Inputs The skill knows nothing about git or bug reports. It reads: - **plan files**, through `fragility-common/lib/plan-format.ts`; - **`targets.json`** in the run directory, written by system-history; - **flags**, for bug context (`--source`, `--note`). ## Scripts - `doctor.ts`: checks credentials, the five project permissions, that the Pod option is on the Task create screen, and the run link. A `run-link` warning means set `FRAGILITY_RUN_URL` or pass `--run-url` to the upsert. - `jira-match.ts --run <id> [--systems a,b --files x,y]`: fetches every `refactor-findings` ticket into `jira-findings.json`. Then it matches them against `targets.json`, or the given systems and files, by fingerprint, system, nesting, and shared files, into `jira-matches.json`. Run it before investigating, so you don't re-plan ticketed work. - `jira-sighting.ts --run <id> --key KEY --system <system> [--apply]`: records that a ticketed system came up again. It comments at most once per cooldown, or once every four cooldowns on a declined ticket. With `--source <url> --source-title "<title>" --note "<how>"` (a bug traced to the ticket) it always comments once. It refuses tickets resolved as fixed: that is a recurrence, so upsert a plan instead. - `jira-upsert.ts --plan <file> [--apply]`: turns a finished plan into Jira state. It is a dry run without `--apply`, and idempotent on re-run: - no ticket: create one with the Pod, the label, the plan attached, and the run link; - open ticket: a sighting, plus a comment and the updated plan once per cooldown. Bug-triggered plans always comment once; - fixed ticket: a new recurrence ticket linked to the old one; - declined ticket: a sighting, a rare comment, and never re-filed. It refuses unfinished plans (`TODO(agent)` markers, missing confidence), `low`-confidence creates, and creates past `maxNewTicketsPerRun` unless `--over-cap`. Use `--duplicate-of KEY` when you judge a plan to duplicate a ticket under a different fingerprint. Each applied action appends to the run's `jira-results.json`, and the upsert writes the ticket key back into the plan.