Analyze mini
Portable AI-engineering skills for Claude Code, Codex, and coding agents: bounded scope, mini-specs, vertical slices, verification, ship gates, and handoff.
npx -y skills add tmusser/ai-engineering-skills --skill analyze-miniAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 3 stars3 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Risk-triggered consistency checkpoint that detects when task artifacts, assumptions, or verification no longer support a safe build decision.
SKILL.md
5.1 KB, as published. Nobody here has run it
Analyze Mini
Purpose
Prevent stale or contradictory reasoning from crossing into build-one without turning analysis into a routine gate.
analyze-mini is conditional. The cheap safeguard is an eligibility check over artifacts already being read for build or handoff. Run the full skill only when that check finds a meaningful trigger.
When to use
Run a full analysis when one or more of these are true:
HANDOFF.mdmarks analysisREQUIREDorSTALE- task-defining inputs changed since the last
FRESHanalysis:SPEC.md,PLAN.md,TODO.md, scope-freeze notes, selected task, or a decision/assumption that changes implementation - the selected task cannot be traced cleanly from acceptance criterion to implementation slice to verification
- an unresolved decision or assumption could materially change implementation shape
- compatibility seams, data/metric semantics, or cross-boundary behavior create interpretation risk
- failed verification caused the implementation strategy or plan to change
- reconciling a stale handoff changed the live task state
Do not run a full analysis merely because the skill is installed, because a handoff is being written, or because no prior analysis exists.
Missing prior analysis is not, by itself, a trigger.
Staleness is event-driven, not time-driven. A checkpoint does not expire just because time passed.
Inputs
Use the smallest relevant set:
- Selected task
SPEC.mdPLAN.md/TODO.mdwhen they exist- Scope-freeze boundary
- Relevant
HANDOFF.mdstate when resuming VERIFY.mdwhen a failure or changed strategy is part of the trigger- Only the source files needed to resolve a specific contradiction
Passive eligibility check
Before invoking the full skill, reuse artifacts already loaded for build-one or handoff and ask:
- Did a task-defining input change since the last trustworthy checkpoint?
- Is there an unresolved decision or assumption that could change the implementation?
- Can the selected task still map cleanly from acceptance criterion to verification without contradicting a non-goal, spec ceiling, or scope boundary?
- Did a failed attempt or resume reconciliation create a new implementation uncertainty?
If all answers are no, use NOT_NEEDED. Do not perform a broader repo scan just to prove that analysis was skipped.
Checkpoint states
Analysis checkpoint: FRESH | NOT_NEEDED | STALE | REQUIRED
FRESH— a full analysis was run against the current task-defining inputs.NOT_NEEDED— the cheap eligibility check found no trigger for a full analysis.STALE— a priorFRESHanalysis exists, but a task-defining input changed.REQUIRED— a current trigger exists and a full analysis must run before implementation continues.
FRESH and NOT_NEEDED permit implementation only when no separate blocker exists.
Workflow
- Name the trigger. Do not start with a generic repo review.
- Identify the smallest artifact delta that could affect the selected task.
- Map acceptance criterion -> selected task -> verification evidence.
- Check non-goals, spec ceiling, compatibility seams, and scope-freeze boundaries for contradictions.
- Check decisions and assumptions that can materially change implementation shape.
- Separate contradictions from unknowns; do not invent certainty to make the artifacts agree.
- Make an implementation readiness call:
READYorBLOCKED. - Emit a compact checkpoint:
ANALYSIS CHECKPOINT
State: FRESH
Readiness: READY | BLOCKED
Trigger: ...
Inputs reviewed: ...
Finding: ...
Criterion -> task -> proof: ...
Re-run when: ...
- If a handoff is being updated, carry the checkpoint state and reason into
HANDOFF.md. Otherwise return the compact checkpoint inline. - Do not create
ANALYZE.mdfor a clean checkpoint. Create or update it only when material findings need durable standalone review or the user explicitly asks for it.
Outputs
- Compact analysis checkpoint
READYorBLOCKEDimplementation call when full analysis runs- Material contradictions, unknowns, or verification gaps
- Optional
ANALYZE.mdonly when durable standalone findings are justified
Stop conditions
NOT_NEEDEDis justified by the cheap eligibility check, or- the full analysis establishes a current
FRESHcheckpoint andREADY, or - a contradiction or unresolved decision makes readiness
BLOCKED
Anti-patterns
- Running
analyze-minibefore everybuild-onemerely because it is available. - Treating a missing prior analysis as analysis debt by default.
- Re-reading the whole repository to decide whether analysis is needed.
- Creating
ANALYZE.mdafter every clean consistency check. - Expiring a valid checkpoint because of elapsed time rather than a state change.
- Keeping a checkpoint
FRESHafter task-defining inputs changed. - Starting implementation while a
REQUIREDorSTALEcheckpoint is unresolved. - Adding complexity because analysis surfaced an interesting adjacent improvement.