Claude
Skill
judgment-day
Trigger: judgment day, dual review, adversarial review, juzgar. Run explicit blind dual review with at most two scoped fix/re-judgment rounds.
Virus-scanned
Reviewed automatically before listing.
Download
gentleman-programming-gentle-ai-internal_assets_skills_judgment-day-d2e3443.zip · 3 KB
Install
skills CLI
npx skills add https://github.com/Gentleman-Programming/gentle-ai/tree/main/internal/assets/skills/judgment-day
Claude Code
claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install gentleman-programming-gentle-ai@llmmart
Git
git clone https://github.com/Gentleman-Programming/gentle-ai.git
The skills CLI installs just this skill, for any of its supported agents. Claude Code installs the whole gentleman-programming/gentle-ai collection as a plugin from our marketplace. Git is the plain clone.
Skill manifest
Activation Contract
Load only when the user explicitly requests Judgment Day or equivalent dual/adversarial review for a concrete target. Judgment Day is a standalone developer tool: judges run whenever asked, on any runtime, and need no review transaction, runtime identity, or delivery-receipt machinery to start. It replaces ordinary 4R as the adversarial method for that target; never run both.
Hard Rules
- Resolve matching project skills before starting and pass the same paths to both judges (
jd-judge-a,jd-judge-b) and the fix actor (jd-fix-agent). - Build one complete immutable target, then launch two blind read-only judges in parallel with identical scope and criteria.
- Each judge returns one neutral findings result and terminates. Wait for both; never accept a partial judgment.
- Never launch
review-refuter; two-judge agreement is the corroboration mechanism. - Only the parent orchestrator merges/persists findings, launches the fix actor (
jd-fix-agent), and launches scoped re-judgment. - Fix only severe findings confirmed by both judges. WARNING/SUGGESTION rows remain
info. - Permit at most two fix rounds and two scoped re-judgments. Re-judgment sees only the frozen ledger plus fix delta and may record fix-caused defects.
- The only terminal verdicts are
APPROVED | ESCALATED; never reset or extend an exhausted round budget. - A judgment issues no receipt and carries no delivery authority: it satisfies no commit, push, PR, or release gate. When the caller explicitly wants ordinary negotiated review for the same target, run it as its own step; neither outcome authorizes delivery, which remains under ordinary repository policy.
Decision Gates
| Condition | Action |
|---|---|
| Target unclear | Ask one scope question and stop. |
| Both judges confirm severe finding | Ask before round-one correction; then use the bounded fix actor. |
| One judge reports it | Record suspect; do not auto-fix. |
| Judges contradict | Escalate for explicit human decision. |
| Scoped re-judgment fails before round two | Parent may launch the final bounded fix round. |
| Any issue remains after round two | Escalate and stop. |
Execution Steps
- Build the complete immutable target and freeze the scope both judges will inspect.
- Launch both read-only judges in parallel (
jd-judge-a,jd-judge-b) against the same immutable target. - Merge findings into the frozen ledger and persist it through the selected artifact store.
- Ask before round-one correction; run the fix actor (
jd-fix-agent) only for confirmed severe IDs. - Run both judges again (
jd-judge-a,jd-judge-b) only over the frozen ledger plus immutable fix delta. - Repeat once at most, then run independent final verification and return the terminal verdict.
Output Contract
Return target identity, round, confirmed/suspect/contradiction/INFO counts, correction work units, scoped re-judgment result, artifact references, skill resolution, and exactly one final JUDGMENT: APPROVED ✅ or JUDGMENT: ESCALATED ⚠️.
References
- references/prompts-and-formats.md — compact judge/fix prompts and verdict shape.
- ../_shared/review-ledger-contract.md — optional ordinary negotiated-review context: consult it only when the caller explicitly requests that lifecycle; never required to run judges and never delivery authorization.
Files (gentle-ai)
-
references
-
prompts-and-formats.md 3 KB
# Judgment Day Prompts and Formats ## Judge Prompt ```markdown You are blind Judge {A|B} in explicit Judgment Day mode. Target: {immutable target identity and exact paths} Skills to load: {resolved SKILL.md paths} Criteria: correctness, edge cases, error handling, performance, security, and project conventions. Run one exhaustive read-only sweep. Do not edit, delegate, refute, or inspect unrelated scope. If scoped re-judging, read ONLY the frozen ledger and immutable fix delta; record any fix-caused defect with proof. Return one JSON object and no prose, using exactly this native result shape: {"findings":[{"location":"path:line or path:start-end","severity":"CRITICAL","claim":"observable incorrect behavior","evidence_class":"deterministic","causal_disposition":"introduced","proof_refs":["concrete proof"]}],"evidence":["what was inspected"]} This is a judgment-day judge result, not a `gentle-ai review capture-result` lens artifact. Judgment day selects no lenses and records your work as a judge proof, so your result carries no bound artifact subject and no inspection envelope. The only allowed top-level fields are `findings` and `evidence`, and the only allowed finding fields are `location`, `severity`, `claim`, `evidence_class`, `causal_disposition`, and `proof_refs`. Never emit `summary`, `skill_resolution`, or any other unknown field. Keep orchestration metadata outside the native result JSON; `evidence` contains only genuine inspection evidence. Return `{"findings":[],"evidence":["what was inspected"]}` when clean, then terminate. ``` > This shape governs judgment-day judges only. An ordinary bounded review lens is a different artifact: it is captured with `gentle-ai review capture-result`, and its result must additionally echo the binding's top-level `subject_hash` and a completed `inspection` envelope, or admission refuses it. `gentle-ai review schema reviewer` emits that schema with a working example. The two shapes are not interchangeable, and neither one is a relaxation of the other. ## Fix Actor Prompt ```markdown You are the bounded Judgment Day fix actor. Confirmed severe ledger IDs: {table} Skills to load: {resolved SKILL.md paths} Apply only confirmed fixes as atomic work units. For each unit, record focused test result, runtime evidence or justified N/A, and rollback boundary. Never review, add findings, refactor unrelated code, or launch another actor. Mark addressed IDs fixed and return control to the parent orchestrator for scoped re-judgment. End with: Skill Resolution: {paths-injected|fallback-registry|fallback-path|none} — {details} ``` ## Verdict Shape ```yaml target_identity: <sha256> round: 1 | 2 confirmed: [] suspect: [] contradictions: [] info: [] fix_work_units: [] scoped_rejudgment: approved | escalated | not_run terminal_state: approved | escalated skill_resolution: <value> ``` Warnings and suggestions are informational. After the second scoped re-judgment, remaining severe findings require `escalated`; no third round exists.
-
-
SKILL.md 3.6 KB
--- name: judgment-day description: "Trigger: judgment day, dual review, adversarial review, juzgar. Run explicit blind dual review with at most two scoped fix/re-judgment rounds." license: Apache-2.0 metadata: author: gentleman-programming version: "1.7" --- ## Activation Contract Load only when the user explicitly requests Judgment Day or equivalent dual/adversarial review for a concrete target. Judgment Day is a standalone developer tool: judges run whenever asked, on any runtime, and need no review transaction, runtime identity, or delivery-receipt machinery to start. It replaces ordinary 4R as the adversarial method for that target; never run both. ## Hard Rules - Resolve matching project skills before starting and pass the same paths to both judges (`jd-judge-a`, `jd-judge-b`) and the fix actor (`jd-fix-agent`). - Build one complete immutable target, then launch two blind read-only judges in parallel with identical scope and criteria. - Each judge returns one neutral findings result and terminates. Wait for both; never accept a partial judgment. - Never launch `review-refuter`; two-judge agreement is the corroboration mechanism. - Only the parent orchestrator merges/persists findings, launches the fix actor (`jd-fix-agent`), and launches scoped re-judgment. - Fix only severe findings confirmed by both judges. WARNING/SUGGESTION rows remain `info`. - Permit at most two fix rounds and two scoped re-judgments. Re-judgment sees only the frozen ledger plus fix delta and may record fix-caused defects. - The only terminal verdicts are `APPROVED | ESCALATED`; never reset or extend an exhausted round budget. - A judgment issues no receipt and carries no delivery authority: it satisfies no commit, push, PR, or release gate. When the caller explicitly wants ordinary negotiated review for the same target, run it as its own step; neither outcome authorizes delivery, which remains under ordinary repository policy. ## Decision Gates | Condition | Action | |---|---| | Target unclear | Ask one scope question and stop. | | Both judges confirm severe finding | Ask before round-one correction; then use the bounded fix actor. | | One judge reports it | Record suspect; do not auto-fix. | | Judges contradict | Escalate for explicit human decision. | | Scoped re-judgment fails before round two | Parent may launch the final bounded fix round. | | Any issue remains after round two | Escalate and stop. | ## Execution Steps 1. Build the complete immutable target and freeze the scope both judges will inspect. 2. Launch both read-only judges in parallel (`jd-judge-a`, `jd-judge-b`) against the same immutable target. 3. Merge findings into the frozen ledger and persist it through the selected artifact store. 4. Ask before round-one correction; run the fix actor (`jd-fix-agent`) only for confirmed severe IDs. 5. Run both judges again (`jd-judge-a`, `jd-judge-b`) only over the frozen ledger plus immutable fix delta. 6. Repeat once at most, then run independent final verification and return the terminal verdict. ## Output Contract Return target identity, round, confirmed/suspect/contradiction/INFO counts, correction work units, scoped re-judgment result, artifact references, skill resolution, and exactly one final `JUDGMENT: APPROVED ✅` or `JUDGMENT: ESCALATED ⚠️`. ## References - [references/prompts-and-formats.md](references/prompts-and-formats.md) — compact judge/fix prompts and verdict shape. - [../_shared/review-ledger-contract.md](../_shared/review-ledger-contract.md) — optional ordinary negotiated-review context: consult it only when the caller explicitly requests that lifecycle; never required to run judges and never delivery authorization.
Comments (0)
Sign in to join the conversation.
Reviews (0)
No reviews yet.
No comments yet.