ilandahan/AID / skills/reflection
reflection
VerifiedAutomatic quality check on significant outputs via isolated sub-agent. Spawns reflection-agent for objective evaluation without conversation context bias.
| File | Kind | Size |
|---|---|---|
| SKILL.extended.md | markdown | 13.1 KB |
| SKILL.md | skill md | 4.8 KB |
| criteria/phase-0-discovery.yaml | config | 2.5 KB |
| criteria/phase-1-prd.yaml | config | 2.5 KB |
| criteria/phase-2-tech-spec.yaml | config | 2.6 KB |
| criteria/phase-3-impl-plan.yaml | config | 9.0 KB |
| criteria/phase-4-development.yaml | config | 3.4 KB |
| criteria/phase-4a-code-design.yaml | config | 3.9 KB |
| criteria/phase-4b-code-function.yaml | config | 5.0 KB |
| criteria/phase-5-qa-ship.yaml | config | 3.7 KB |
criteria/phase-4a-code-design.yaml
View on GitHub# Phase 4a: Code Design - Reflection Criteria (Autoresearch AR-1 / DESIGN pass)
# Pre-TDD design/quality reflection pass. Judge-only scoring.
# Generalized from the Waze rubric dimensions E/F (structure, loud failures) + B.
phase: "4a"
name: "Code Design (AR-1)"
description: "Pre-TDD design & quality reflection: structural health, loud failures, naming, traceability. NO interface/public-API/data-shape changes."
criteria:
- id: structural_health
name: "Structural Health"
weight: 3
checks:
- "No copy-paste duplication; repeated logic is factored into one place?"
- "Functions/modules are decomposed to a single clear responsibility (no god-functions)?"
- "Magic numbers and literals are centralised into named constants/config, not scattered?"
- "Module boundaries are clean; no leaking of internal concerns across layers?"
scoring:
excellent: "Well-decomposed, DRY, constants centralised, clean module boundaries"
acceptable: "Mostly clean with minor duplication or a few stray literals"
poor: "Significant duplication, god-functions, scattered magic numbers, or tangled boundaries"
- id: loudness_of_failures
name: "Loudness of Failures"
weight: 3
checks:
- "Every catch block surfaces the failure (banner/throw/log) or is justified by an inline comment?"
- "Every '|| DEFAULT' fallback is intentional and surfaced/justified, not silently masking an error?"
- "Every '?? DEFAULT' fallback is intentional and surfaced/justified?"
- "No swallowed promises (.catch without rethrow/log, fire-and-forget async)?"
scoring:
excellent: "All failure paths are loud or explicitly justified; nothing swallowed silently"
acceptable: "Failures mostly surfaced; one or two defaults lack justification comments"
poor: "Silent catches, unexplained '||'/'??' defaults, or swallowed promises mask errors"
- id: naming_clarity
name: "Naming Clarity"
weight: 2
checks:
- "Names reveal intent (what the thing is/does), not just its type?"
- "No misleading names that contradict behaviour?"
- "No cryptic abbreviations or single-letter names outside tight loops?"
scoring:
excellent: "Names are intent-revealing, accurate, and unabbreviated"
acceptable: "Mostly clear names with a few abbreviations or vague identifiers"
poor: "Misleading, cryptic, or abbreviation-heavy names obscure intent"
- id: docs_traceability
name: "Docs & Traceability"
weight: 2
checks:
- "File/module headers explain purpose and ownership?"
- "Non-obvious decisions carry WHY comments (rationale, not restating the code)?"
- "Code links back to the spec/task it implements (task id, ticket, or spec reference)?"
scoring:
excellent: "Headers present, WHY comments on non-obvious logic, clear links back to spec/task"
acceptable: "Some documentation; partial traceability to spec/task"
poor: "No headers, no WHY comments, no traceability to the originating spec/task"
# Aligned with kpi_target (>= 9.5) — design pass must clear a high bar before TDD.
threshold: 8.0
max_iterations: 3
notes:
pass_type: "PRE-TDD design pass (AR-1). Judge-only: no running tests exist yet at this stage."
hard_constraints:
- "Exported signatures / public API must NOT change."
- "Data shapes / contracts must NOT change (keeps TDD from chasing a moving target)."
- "One focused edit per iteration; edit only source changed in this task."
- "No new dependencies, no framework migration, no plaintext secrets."
- "Never auto-commit; never destructive git. Validation only."
borderline: "Review loudness_of_failures justification comments and magic-number centralisation."
improvement_focus:
- "Surface or justify every catch / '||' / '??' default"
- "Factor out duplication and centralise magic numbers"
- "Make names intent-revealing"
- "Add file headers, WHY comments, and spec/task links"
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |