critique

Solid

The critique node of the canonical workflow — run two adversarial critics in parallel (implementer lens + user lens) against a freshly-planned spec, write their findings to .oh/tasks/<slug>/critique.md, and hand off to the approve gate. This is the EVIDENCE half of the critique→approve pair; it produces findings, it does not decide (the /approve gate decides). Runs on local artifacts only, before any GitHub-side state. TRIGGER when: a PRD/plan exists and needs adversarial review before the commitment gate; "critique this spec", "run the critics on <slug>", "review prd before we build".

AI & Automation 38 stars 3 forks Updated today Apache-2.0

Install

View on GitHub

Quality Score: 83/100

Stars 20%
53
Recency 20%
100
Frontmatter 20%
70
Documentation 15%
100
Issue Health 10%
50
License 10%
100
Description 5%
100

Skill Content

# Critique — two adversarial critics before commitment The **critique** node of the canonical workflow (`AGENTS.md § The Workflow`), composed by `/spec critique`. It runs the short adversarial feedback loop on a planned spec *before* anything is committed, and writes the `critique.md` artifact the `approve` gate then decides on. **Core principle: surface risk while the spec is still the cheapest thing to revise.** Two critics with **different framings** review `.oh/tasks/<slug>/prd.md` in parallel — symmetric critics waste context. Every finding is SEVERITY-tagged so the downstream `/approve` gate can act on it mechanically. This node produces evidence; it does not gate (that is `/approve`). Like the gate, it touches only local artifacts — no GitHub-side state exists yet (critic-before-commitment). Formalizes `/ship-spec` Stage 3 as a reusable node. --- ## Inputs | Arg | Meaning | |-----|---------| | `<slug>` | The task slug — the critics read `.oh/tasks/<slug>/prd.md`; output is written to `.oh/tasks/<slug>/critique.md`. Required. | If `.oh/tasks/<slug>/prd.md` is absent there is nothing to critique — print an error pointing at `/spec plan` and emit **no** `STATUS:` token (honest exits: a missing spec is a failure, not a clean critique). --- ## Run the critics (parallel) Launch **two `Agent` tool calls in a single message** (parallel execution) with `subagent_type: "critic"`. Both also receive the protected-paths cross-check: read `.claude/protected-paths.txt` and f...

Details

Author
mifunedev
Repository
mifunedev/openharness
Created
5 months ago
Last Updated
today
Language
TypeScript
License
Apache-2.0

Similar Skills

Semantically similar based on skill content — not just same category

AI & Automation Solid

approve

The approve (approve|deny) decision gate of the canonical workflow — read a spec's critique findings and decide go/no-go BEFORE any GitHub-side state exists, then emit a single APPROVED/DENIED verdict. Auto-gates on critic SEVERITY (any unmitigated high → deny), with an optional human override for borderline findings. This is the decision half of the critique→approve pair; the critics that produce the findings are the separate `/critique` node. TRIGGER when: a plan/PRD has been critiqued and needs a commitment decision before implementation; "approve this spec", "should we build <slug>", "gate <slug> on its critique".

38 Updated today
mifunedev
AI & Automation Listed

critique

You are a rigorous, adversarial critic. Your job is to find real problems — in code, strategy, logic, design, or any other domain. You serve as both an engineering gate and a… Use for /critique requests.

2 Updated 3 days ago
amurshak
AI & Automation Listed

devils-advocate

Adversarially critique a proposal by generating alternatives. Dispatches 3 parallel critics (pragmatist, paranoid, architect lenses) — each invents one alternative approach — then a synthesis step ranks all 4 options and recommends the top choice. When the proposal was authored by someone other than the agent (inherited plan, someone else's PR, external review), a 4th steelman critic runs in the same parallel wave and strengthens the original first, so it is judged at its strongest rather than its weakest. Use when a plan, fix, scoping, decomposition, or named recommendation will drive decisions and you want structured alternative-generation before committing. Complements /shadow-verify — that skill re-derives factual claims; this one critiques whether the chosen approach itself is best.

54 Updated today
griffinwork40