cross-spec-consistency-check
Skill2 starsChanged 3 days agoMITView on GitHub
Pre-authoring conflict detection across the specs in a build's source-grounding set — API naming drift, double authority, schema drift, authority overlap. Triggers when a build run has 2+ source specs to check before authoring (build specify); NOT for a greenfield spec with no source set.
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 3 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 17
Read and follow framework/work-principle.md before taking up work.
- Distinctive: line 132
Retry: after an existing-spec patch or council trigger
- Distinctive: line 142
DO NOT: check all existing specs in the repo against each other (combinatorial explosion).
- Distinctive: line 143
DO NOT: write without severity triage.
≈ “Common” is an estimate: at least 1 in 5 of the 308,232 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Read and follow framework/work-principle.md before taking up work.
Resolve this reference against the framework root; retain the consumer CWD.
Skill: cross-spec-consistency-check
Purpose
Conflict detection between specs in the source-grounding set, before a NEW spec is written that builds on inconsistent predecessors. Prevents double authority and drift inheritance from leaking into the NEW spec output.
Executed in the build workflow BEFORE the interview step. Block on high-severity conflicts.
Who executes
Buddy as orchestrator. Optionally a board-impact sub-agent for cross-spec diff. Triage decisions are substance, not delegated.
Input
| Parameter | Type | Required | Description |
| build_id | string | yes | Current build-workflow run identifier |
| parent_task | int | yes | Build task ID |
| source_grounding_set | list[str] | yes | List of spec paths read as input for the current authoring |
Source-grounding set: the project's own spec-consumed list for the current build scope, plus any existing outputs from earlier build iterations that may compete with the current build in substance. Concrete sources are a consumer convention (e.g., spec-map index, feature map, plan authority).
Procedure
Step 1: Assemble the source-grounding set
Buddy + author identify which spec files are relevant as input for the current build. Typical sources:
- Spec index with consumed/consumer columns
- Feature or capability map
- Outputs of earlier build iterations (NEW specs that already exist)
- Architecture authority stub if present
Step 2: Conflict identification
Per spec pair (i, j) in the source-grounding set: identify the diff. Four pattern classes:
- API naming drift: the same identifier in two specs with a different signature or semantics
- Concept double authority: the same concept in two specs with different descriptions
- Schema drift: the same data structure in two specs with different definitions
- Authority overlap: both specs declare authority for the same aspect
Methods: Buddy direct, manual, or sub-agent delegation. Script-based pattern detection (same identifier via grep + AST comparison) is a phase-2 improvement, not mandatory.
Step 3: Write drift-list-cross.md
Output path: docs/build/<build-id>/drift-list-cross.md
Format:
# Cross-Spec-Drift-List
**Build:** {build-id}
**Generated:** YYYY-MM-DD
**Source-Grounding-Set:** [list of specs]
**Total Conflicts:** {N}
## Conflicts
### CSC-{build-id}-001
- **Specs involved:** {spec-a}.md §X.Y ↔ {spec-b}.md §A.B
- **Description:** {conflict description}
- **Severity:** high | medium | low
- **Resolution-Path:** patched-existing-spec | decision-in-new-spec | council-triggered
- **Resolved-by:** {commit-sha or PENDING}
- **Resolution-Note:** {rationale}Step 4: Severity triage
| Severity | Definition | Block behaviour |
| high | Conflict blocks authoring (the author cannot decide which authority applies) | Block — must be resolved before authoring |
| medium | Conflict resolvable in the NEW spec (decision lock there), source patch optional | Warn |
| low | Naming drift, cosmetic | Warn |
Step 5: Resolution
Per high-severity conflict, three resolution paths:
- (a) patched-existing-spec: the existing spec is patched immediately (co-evolution)
- (b) decision-in-new-spec: the NEW spec declares explicitly which authority it overrides (§Authority-Conflicts block in the spec body)
- (c) council-triggered: Council with 3+ members, ADR output
Severity medium/low: resolution during the subsequent authoring step or post-Specify.
Contract
INPUT
- Required: build_id, parent_task, source_grounding_set
- Context: source-grounding-set files (read-only)
OUTPUT
DELIVERS:
docs/build/<build-id>/drift-list-cross.md- Resolution commits for high severity (before the next step)
DOES NOT DELIVER:
- Spec authoring (that is
spec_authoring) - Source-spec reduction (that is
source_spec_reduce) - ADR writing (manual when council-triggered)
ENABLES:
- Clear authority layers during spec authoring
- Cross-build-iteration consistency (sequential builds build on each other without drift inheritance)
DONE
- drift-list-cross.md written
- High-severity conflicts resolved
- Medium/low documented with resolution path
FAIL
- Retry: after an existing-spec patch or council trigger
- Escalate: conflict outside the four pattern classes -> user escalation
- Abort: > 5 high-severity conflicts -> stop, mandatory substance audit (signal of a structural problem)
Skip eligibility
Skip-eligible when the current build-workflow run has no source-grounding set (greenfield spec without an existing spec predecessor or without intersecting specs from earlier builds). Skip reason as a one-sentence note in the --skip invocation.
Anti-patterns
- DO NOT check all existing specs in the repo against each other (combinatorial explosion). INSTEAD: only the source-grounding set of the current build. Why: conflicts outside the current scope are not build-relevant.
- DO NOT write without severity triage. INSTEAD: an explicit classification per conflict. Why: without severity, the file is useless for a block/warn decision.
- DO NOT misuse this as a pre-project audit for all specs. INSTEAD: incremental per build. Why: bulk audit waves have historically failed under load.
Enforcement
workflows/runbooks/build/workflow.yamlstepcross-spec-consistency-check— pre-interview, on_fail: block- The consumer project defines its own spec-lifecycle policy authority (e.g., ADR) for when skip eligibility applies and when source-reduction end state is reached. This skill is the generic mechanism; project-specific triggers are a consumer convention.
In this repository85 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- adversary-test-planSkill
- api-and-interface-designSkill
- architecture-coherence-reviewSkill
- bedrock-drillSkill
- cavemanSkill
- code-review-boardSkill
- consistency-checkSkill
- councilSkill
- cross-spec-consistency-checkSkill
- deprecation-and-migrationSkill
- documentation-and-adrsSkill
- frameSkill
- frontend-design-ttySkill
- frontend-ui-engineeringSkill
- get-api-docsSkill
- impl-plan-reviewSkill
- improve-codebase-architectureSkill
- knowledge-captureSkill
- knowledge-processorSkill
- pre-build-spec-auditSkill
- retroactive-spec-updateSkill
- root-cause-fixSkill
- scopingSkill
- sectional-deep-reviewSkill
- security-and-hardeningSkill
- session-diagnosisSkill
- shipping-and-launchSkill
- show-open-tasksSkill
- source-spec-reduceSkill
- spec-authoringSkill
- spec-boardSkill
- task-creationSkill
- task-status-updateSkill
- testingSkill
- transparency-headerSkill
- ux-authoringSkill
- workflow-routerSkill
- youtube-subtitlesSkill
- adversary-test-planSkill
- api-and-interface-designSkill
- architecture-coherence-reviewSkill
- bedrock-drillSkill
- cavemanSkill
- close-retroSkill
- code-review-boardSkill
- consistency-checkSkill
- convergence-loopSkill
- councilSkill
- deprecation-and-migrationSkill
- documentation-and-adrsSkill
- frameSkill
- frontend-design-ttySkill
- frontend-ui-engineeringSkill
- get-api-docsSkill
- impl-plan-reviewSkill
- improve-codebase-architectureSkill
- knowledge-captureSkill
- knowledge-processorSkill
- The first 60 shown.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

