원클릭으로
planner-refine
Targeted fix of validate_plan findings — re-elaboration, duplicate resolution, dependency corrections
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
메뉴
Targeted fix of validate_plan findings — re-elaboration, duplicate resolution, dependency corrections
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
| name | planner-refine |
| categories | ["planner"] |
| description | Targeted fix of validate_plan findings — re-elaboration, duplicate resolution, dependency corrections |
| hooks | {"PreToolUse":[{"matcher":"*","hooks":[{"type":"command","command":"echo '[SKILL: planner-refine] Refining plan...'","once":true}]}]} |
Targeted repair of validate_plan findings. Loads validation.json and repairs each
finding type: re-elaborates failed WPs, resolves duplicate deliverable ownership, and
fixes dependency reference errors. Sizing violations are escalated as CRITICAL. Writes
corrected artifacts back to the output directory so validate_plan can re-run.
The recipe runs this skill with retries: 2 — up to 3 total attempts (1 initial + 2
retries) before escalation.
validate_plan returns verdict: failvalidation.json{{AUTOSKILLIT_TEMP}}/planner/run-YYYYMMDD-HHMMSS)NEVER:
ALWAYS:
validation.json before reading any artifactwp_manifest.json and wp_index.json whenever WP structure changesrefinement_complete and issues_fixed output tokensRead $1. Extract the findings array (contains only error-severity findings as structured
dicts). Extract the message field from each finding for classification. Group by type:
WP .* has status 'failed'WP .* has \d+ deliverablesDeliverable '.*' claimed by multiple WPsWP .* depends on unknown WPPhase .* has no assignments or Assignment .* has no work packagesWP .* has malformed idCycle detected among WPsFile '.*' touched by multiple WPs — these appear in the warnings array, not findings. No action needed; skip if encountered in findings.{$2}/work_packages/wp_manifest.json, {$2}/work_packages/wp_index.json{id}_result.json files for WPs mentioned in the findingsFailed WPs — re-elaborate:
wp_manifest.json (provides name, scope,
estimated_files)model: "sonnet" per failed WP. Provide: WP name, scope,
estimated_files, and the relevant portion of wp_index.json for context{$2}/work_packages/{id}_result.jsonwp_index.jsonwp_manifest.json from failed to doneSizing violations — escalate:
WP .* has \d+ deliverables indicate WPs outside the 1–5 deliverable
sizing bound. Cannot be auto-corrected — the implementation recipe handles re-splitting
downstream.CRITICAL: Cannot auto-fix sizing violation:
- {finding text}
Manual review of WP deliverable allocation required.
Write this to stdout. Do NOT attempt WP splitting or merging.
Duplicate deliverables — resolve ownership:
deliverables (keep in files_touched)deliverables: [], that WP is an orphan. For each orphan:
files_touched entries as deliverables to the most relevant owner
WP (the one with the most scope overlap), selecting at most
DELIVERABLE_BOUNDS[1] - len(owner.deliverables) entries. Any remaining entries stay
in files_touched only.technical_steps and acceptance_criteria into the owner WPdepends_on referenceswp_manifest.json and wp_index.json accordingly
This is a deduplication side-effect resolved in the same step, not a sizing violation.
Deliverable count bounds are defined in schema.py::DELIVERABLE_BOUNDS._result.json for the affected WPsDependency reference errors — fix broken dep IDs:
WP X depends on unknown WP Y finding:
wp_index.json for a WP with a similar name or scope to the missing Y
(it may have been renamed or split)depends_on in {$2}/work_packages/{X}_result.jsondepends_ondep_graph.json exists, update it to reflect corrected dependency IDsMissing assignments/WPs — escalate:
CRITICAL: Cannot auto-fix missing structural elements:
- {finding text}
Manual intervention required before validate_plan can pass.
Write this to stdout. Do NOT attempt structural creation.
Malformed WP IDs — escalate:
WP .* has malformed id (expected PX-AY-WPZ) indicate a corrupted
_result.json or wp_manifest.json. Cannot be auto-corrected without understanding the
intended ID.CRITICAL: Cannot auto-fix malformed WP ID:
- {finding text}
Manual inspection of wp_manifest.json required.
Write this to stdout. Do NOT attempt ID renaming.
DAG cycles — escalate:
Cycle detected among WPs: ... indicate a circular dependency in the
WP graph. Resolving cycles requires semantic understanding of the plan structure.CRITICAL: Cannot auto-fix DAG cycle:
- {finding text}
Manual restructuring of depends_on relationships required.
Write this to stdout. Do NOT attempt cycle-breaking.
Write all modified files back atomically (read current → apply change → write). Modified
files may include: _result.json files, wp_manifest.json, wp_index.json,
dep_graph.json.
Note: Combined documents (
combined_*.json,refined_*.json) are intermediate orchestration artifacts and are NOT updated by this skill. Downstream consumers (validate_plan,compile_plan) read from individual*_result.jsonfiles directly, so stale combined documents do not affect pipeline correctness.
refinement_complete = true
issues_fixed = <N>
N = count of findings addressed from the findings array (failed_wps +
duplicate_deliverables + dep_references). Sizing-violation, missing-element, malformed-ID,
and DAG-cycle findings are excluded from the count (they are escalated as critical, not
fixed). Files-touched overlap findings are in the warnings array and are not actionable.
Analyze issue dependencies and produce a dispatch execution map for parallel orchestration
Execute recipe sessions batch-by-batch for triaged GitHub issues. Reads the triage-issues output manifest, processes each batch sequentially, and launches the appropriate recipe for each issue. Use when user says "process issues", "run issues", or "execute pipeline for issues".
Validate review-decisions audit findings with mandatory intent analysis and seven evidence-gathering rules. Adds docstring-as-contract recognition, deliberate-change detection, test-as-intent-signal, consumer-impact verification, architectural feasibility checks, behavioral simulation, and symmetry-as-design recognition to the standard validation workflow. Use when validating reports from audit-review-decisions specifically.
Generate YAML recipes for .autoskillit/recipes/. Use when user says "make script skill", "generate script", "script a workflow", "write a script", "create a script", "new recipe", "write a pipeline", or when loaded by other skills for script formatting.
Validate audit findings from audit-arch, audit-tests, or audit-cohesion against actual code, git history, and design intent using 9–10 parallel subagents. Removes contested findings, documents exceptions, adjusts severities. Use when user says "validate audit", "validate findings", "validate report", or "check audit results".
Validate audit findings from audit-arch, audit-tests, audit-cohesion, audit-feature-gates, audit-docs, or audit-review-decisions against actual code, git history, and design intent using 9–10 parallel subagents. Removes contested findings, documents exceptions, adjusts severities. Use when user says "validate audit", "validate findings", "validate report", or "check audit results".