refine-ticket — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited refine-ticket (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Given a Jira issue key (e.g. SW-123), this skill reads the ticket via the Atlassian MCP, scores it against the refinement rubric, and loops — auto-resolving gaps from existing automation + docs and asking you (or a source you point it at) for the rest — until the ticket is unambiguous. On your approval it writes a ## Refined Acceptance Criteria block back to the ticket. It does NOT generate tests; run /from-issue after. See ADR-0013.
Tell Claude what you want:
Use the refine-ticket skill on SW-123.
Or to preview without writing to Jira:
Use the refine-ticket skill on SW-123 with dry-run.
During the loop, when asked about a gap you can either answer directly or point the skill at a source ("it's in Confluence page X", a URL, a doc path). The skill ingests it and continues. It does NOT inspect a running app — it's shift-left, so it works even before the feature is built. It stops when the ticket has no open gaps, then asks before writing anything back.
The full procedural workflow is in references/workflow.md. Read that file before executing the skill.
Setup note: the Atlassian MCP must be connected (OAuth) with write scope for the Step 7 description update — defined at project scope in.mcp.json. Its read + write tools are pre-authorized inallowed-toolsabove so the loop doesn't prompt each call. Write-back happens only on your explicit approval (per ADR-0013).
references/workflow.md — the procedural loopreferences/rubric.md — the "bulletproof" checklist (definition of done)references/sources.md — source catalog + user-supplied-source protocolreferences/writeback-template.md — the ## Refined Acceptance Criteria block + idempotent-update ruleThis skill pairs with /from-issue: refine first, then generate. It reuses /from-issue's bucket-classification.md and qa-analysis.md for the rubric's coverage/automatable judgment.
docs/refine-ticket.md — learning guide with a worked exampledocs/adr/0008-custom-skills-pattern.md — why custom skills follow this layout~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.