Add the dev-workflow:sparring advisory skill (0.14.0) - #32
Conversation
Gate-A spec cycle closed. Final pass clean at the derived floor: 0 Blockers
and 0 Majors at pass 3, every earlier Blocker and Major resolved, Minor and
Nit findings collected without iteration. No product file is written by this
commit; implementation is not authorized.
Docs-only change (docs/**.md) — Gate B is N/A per CLAUDE.md §5, and no plugin
path is touched, so invariant 12 does not apply.
cycle at71dccpoc; floor 3 per {docs/superpowers/stories/2026-09-17-sparring-skill-story.md (level 1)}; hook reminder threshold absent
cycle at71dccpoc; Gate-A spec (passes 1-3, gpt-6-astra): Findings 18,10,3. Blockers 0,0,0. Majors 6,4,0.
Gate-A plan cycle closed. Pass 5 is clean at or above the derived floor:
0 Blockers and 0 Majors; every earlier Major was repaired in the plan and
confirmed resolved by the following pass. Pass 5's two Minors are
collected, not iterated. The committed plan is byte-identical to the text
pass 5 reviewed (sha256 f1e84f110f899ecd8de03022eb2aa466978973966413bcee67d24adac45c46d1).
Docs-only change (docs/**.md) — Gate B is N/A per CLAUDE.md §5.
cycle vl584i4v7j; floor 3 per {docs/superpowers/stories/2026-09-17-sparring-skill-story.md (level 1)}; hook reminder threshold absent
cycle vl584i4v7j; Gate-A plan (passes 1-5, gpt-6-astra): Findings 15,10,6,4,2. Blockers 0,0,0,0,0. Majors 6,6,1,1,0.
New skill plugins/dev-workflow/skills/sparring/SKILL.md, copied verbatim from
the spec's §5: an explicitly invoked advisory session (read-only by
instruction, not a sandbox) for investigating, checking agent reports
against the current files, and drafting bounded prompts for a coding agent
to run elsewhere. disable-model-invocation: true stops the model invoking
it on its own; it restricts nothing once running. Loaded by convention, no
manifest key (invariant 6).
docs/sparring-briefing.md's closing section said the pattern was "not a
plugin feature"; this change made that false, so it is replaced with the
spec's §6 text, which records that the promotion happened by decision, not
by a trigger. The skill is named in README.md, AGENTS.md and
docs/architecture.md. dev-workflow 0.13.3 -> 0.14.0.
Evidence — docs/superpowers/stories/2026-09-17-sparring-skill-story.md
Battery: AGENTS.md quality row, exit 0 at fcfef04cd30789a4eb3a6c680babae0a1fdda53d.
Check (counterfactual, spec §7 row 7): grep -c 'Not a plugin feature' and
'trigger, not a reflex' in docs/sparring-briefing.md read 1 and 1 before the change,
0 and 0 after.
cycle 73tpkveiel; floor 3 per {docs/superpowers/stories/2026-09-17-sparring-skill-story.md (level 1)}; hook reminder threshold absent
cycle 73tpkveiel; Gate B (passes 1, gpt-6-astra): Findings 0. Blockers 0. Majors 0.
Gate B pass 1 was one logical pass run as two calls (reviewType spec and
quality, same baseSha dc44dd5 and headSha
fcfef04cd30789a4eb3a6c680babae0a1fdda53d); both returned NO FINDINGS, so the
cycle closed on the zero-finding exit below the floor.
|
Navigate logical layers of code changes, visualize relationships, and explore their blast radius. Warning Review limit reachedYou've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository. Next included review available in 31 minutes. View limit detailsLimit details: You’ve used the included review currently available. Review configuration: ⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Advanced Run ID: 📒 Files selected for processing (2)
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Advanced Run ID: 📒 Files selected for processing (10)
Included review availability: This review used your included allowance. Your plan provides up to 1 included review per hour; 0 remain after this review. 📝 WalkthroughWalkthroughThe pull request adds the explicitly invoked ChangesSparring Skill
Priority: ⬇️ Low Estimated code review effort: 3 (Moderate) | ~20 minutes Change: Feature Sequence Diagram(s)sequenceDiagram
participant User
participant SparringSkill
participant Repository
participant CodingAgent
User->>SparringSkill: Request investigation or advice
SparringSkill->>Repository: Read project guidance and current artifacts
Repository-->>SparringSkill: Provide repository evidence
SparringSkill-->>User: Return findings and, if requested, a bounded prompt
User->>CodingAgent: Provide prompt in a separate coding session
Merge Risk: ⚪ Minimal · up to The advisory skill and its release documentation are consistent, with no concrete merge-blocking issue identified. Its read-only limits are instructions, not a sandbox. Mergeable subject to normal checks. Security Architecture ReviewSecurity architecture risk: 🔵 Low · up to The advisory workflow clearly limits intended actions and separates advice from implementation. Those limits are instructions rather than enforced restrictions. No concrete security defect was established, but execution behavior and exceptional document-write recovery remain unverified. Retained concerns Security review detailsSecurity Blast Radius
Trust Boundaries and Controls
Resilience and Maintainability Implications
Hardening Proposals
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. A rabbit reads the files with care, Comment |
 |
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
…urrence The sparring-skill spec introduced its walkthrough list as "the eight walkthrough scenarios … the last two were added by pass 1" while the list numbers twelve; the second Gate-A repair round added 9-12 without updating the sentence (PR #32, Greptile P2). The sentence now says twelve and which round added which. docs/hardening-log.md records it as the eighth docs-drift occurrence, at 1 prose with no new text. Docs-only change (docs/**.md): Gate B is N/A per CLAUDE.md §5's prose exemption, so there is no Gate-B cycle and no provenance line or skip record.
1 participant
What
A new user-invoked skill,
dev-workflow:sparring: an advisory session in a chat you open for that purpose. It investigates read-only, checks agent reports against the current files, and drafts bounded prompts for a coding agent to run elsewhere. It does not implement, commit, or run review gates. The read-only posture is an instruction, not a sandbox.plugins/dev-workflow/skills/sparring/SKILL.md: new, copied verbatim from spec §5, no manifest key (invariant 6)docs/sparring-briefing.md: the closing section said "not a plugin feature" and is replaced by spec §6README.md,AGENTS.md,docs/architecture.md: the skill is named in each inventory0.13.3→0.14.0, with a CHANGELOG entrySpec:
docs/superpowers/specs/2026-09-17-sparring-skill-design.md· Story:docs/superpowers/stories/2026-09-17-sparring-skill-story.md· Plan:docs/superpowers/plans/2026-09-30-sparring-skill.mdReview record
at71dccpoc, 3 passes, closed (commit 01a511d).vl584i4v7j, 5 passes. Findings 15,10,6,4,2 and Majors 6,6,1,1,0. Closed (commit dc44dd5).73tpkveiel, pass 1. Both the spec and the quality branch returned NO FINDINGS, so the cycle closed on the zero-finding exit.What no check covers
What the skill does in a session is prompt text. No harness runs it, so its behaviour was checked by reading only: 12 walkthrough scenarios plus the 12 prompt-standards items, and none of it was executed.
Summary by CodeRabbit
dev-workflow:sparringskill for repository investigation, checking reports against current files, weighing options, and drafting bounded prompts for a coding agent.docs/SPARRING-PARTNER.md.