skills.rest
carterlasalle
carterlasalleCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

traceability

Maintains trace:v1 markers linking requirements, code, tests, and evidence in TraceLayer repositories.

1|Updated Aug 9, 2026
Search Tags:#requirements#verification#agent workflow#evidence#traceability#trace markers#trace cli
One-click install
npx skills add https://github.com/carterlasalle/scc --skill traceability-carterlasalle
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: traceability
Source: https://github.com/carterlasalle/scc/tree/main/.agents/skills/traceability
Command: npx skills add https://github.com/carterlasalle/scc --skill traceability-carterlasalle

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? AI coding agents lose track of why code exists, which requirement it satisfies, and whether tests actually prove it works. This Skill keeps a repository's requirements, decisions, plans, implementations, tests, and evidence connected through trace:v1 markers and the TraceLayer graph, so every change starts with the right context and finishes with verifiable proof. ## Core Features & Use Cases - Trace-oriented workflow: Enforces a 12-step loop — search the trace graph, load context with trace context, implement, run linked tests, ingest evidence, and pass trace verify --changed before completion. - Marker authoring guidance: Teaches correct trace:v1 marker syntax, placement, and edge semantics (satisfies, verifies, exercises, addresses, supersedes) at meaningful behavioral boundaries only. - Lifecycle and knowledge management: Manages the WORK → REQUIREMENT → DECISION/PLAN → IMPLEMENTATION → TEST → EVIDENCE lifecycle, plus durable knowledge nodes and canonical facts with drift detection. - Use Case: When asked to change refresh-token behavior, the agent runs trace context impl.auth.refresh, edits the code while preserving its stable trace ID, re-runs the linked test, ingests the evidence, and finalizes the work item only after trace verify --changed passes. ## Quick Start Ask the agent to implement a requirement or fix a bug in a TraceLayer-active repository and it will automatically search the trace graph, load context, and verify the change before finishing.

Frequently Asked Questions about traceability

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I add traceability markers to my code?▼

Place a single-line trace:v1 marker directly above the behavioral boundary, such as `# trace:v1 id=impl.auth.refresh work=AUTH-237 satisfies=REQ-AUTH-017`. Only mark meaningful boundaries like API endpoints, business rules, and verification tests — never imports, getters, or generated code.

What is the difference between verifies and exercises in trace markers?▼

The `verifies` edge links a test to the requirement it checks, while `exercises` links the test to the implementation it actually runs. Declare both separately on test markers, then ingest the passing run with `trace evidence ingest` so the proof is observed, not just declared.

Does this skill work in repositories without TraceLayer configured?▼

No. In passive or unconfigured repositories the skill does nothing TraceLayer-specific unless the user activates TraceLayer with `trace init`. Hooks remain passive until activation, after which TraceLayer operates ambiently during development.

What happens when a requirement changes after implementation?▼

Changing a requirement's semantics flags downstream implementations and tests as STALE_REVIEW_REQUIRED and marks prior evidence as historical. You must review the affected artifacts, acknowledge with `trace review`, and produce fresh verification before the work can pass policy again.

Why does trace verify fail before completion?▼

Verification fails when blocking diagnostics exist, such as missing requirement ancestry, unverified tests, unresolved obligations, or open material questions. Every failure carries a rule ID and remediation action — follow it, then re-run `trace verify --changed`.

Can I use TraceLayer through MCP instead of the CLI?▼

Yes. Start the server with `trace mcp` and use eleven tools — orient, brief, next, delta, status, search, context, why, impact, verify, and index — each mirroring its CLI counterpart. Loading context through MCP satisfies the pre-edit gate exactly like the CLI.

Related Skills

Explore More Software Engineering Skills→
carterlasallecarterlasalle
1

traceability

Maintains trace:v1 markers linking requirements, code, tests, and evidence in TraceLayer repositories.

Community
Advanced
tkoguttkogut
1

om-code-review

Reviews diffs, branches, and PRs against correctness, security, and breaking-change standards with severity-ranked findings.

Community
Advanced
Ramprasad4121Ramprasad4121
1

audit-prep-assistant

Prepare codebases for security review with static analysis and documentation.

Community
Advanced
mst-software-vnmst-software-vn
1

setup-matt-pocock-skills

Configures issue tracker, triage labels, and domain doc layout for engineering skills.

Official
Intermediate