skills/trace/SKILL.md
Capture agent-session work records as local JSONL audit evidence. Links a backlog/spec, branch, commits, review verdicts, QA/demo evidence, transcript refs, and shipped ref without storing raw private transcripts. Use when: "trace this work", "write work record", "agent session trace", "journal this delivery", "link transcript evidence". Trigger: /trace, /journal.
npx skillsauth add phrazzld/agent-skills traceInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Capture the work record, not the raw conversation.
/trace is the durable session-lifecycle primitive. It writes sanitized JSONL
records that link a unit of work to its backlog/spec, branch, commits, reviewer
verdicts, QA evidence, demo evidence, transcript refs, and shipped commit/ref.
The default store is .harness-kit/traces/work-records.jsonl.
Trace append/preview commands are usually mechanical and may run direct solo.
When the trace requires substantive evidence classification, provenance
judgment, redaction-risk review, or conflicting artifact synthesis, delegate
on judgment per the shared Roster contract: native subagents by default; add
cross-model critics, roster providers, or sprite lanes (/sprites) only when
they answer a distinct question. See harnesses/shared/AGENTS.md (Roster).
Local lane guidance: Use specialized lanes for evidence classifier, redaction-risk critic, and provenance verifier. Store refs and receipt ids, not raw provider transcripts.
/agent-transcript to render a scoped,
redacted transcript excerpt before attaching it.cargo run --locked -p harness-kit-checks -- trace-record append \
--backlog 056 \
--branch "$(git branch --show-current)" \
--commit "$(git rev-parse --short HEAD)" \
--reviewer-verdict-ref ".harness-kit/traces/delegations.jsonl#<id>" \
--qa-ref ".evidence/qa/<id>.md" \
--demo-ref ".evidence/demo/<id>.gif" \
--transcript-ref ".harness-kit/traces/transcripts/<id>.md"
If transcript evidence is unavailable:
cargo run --locked -p harness-kit-checks -- trace-record append \
--backlog 056 \
--branch "$(git branch --show-current)" \
--commit "$(git rev-parse --short HEAD)" \
--waiver-reason "No safe transcript export available from this harness run."
Smoke test:
cargo run --locked -p harness-kit-checks -- trace-record --self-test
/ship requires a final work record linking the shipped commit to trace
evidence, or an explicit waiver when no transcript or trace artifact exists.
The trace may be a .harness-kit/traces/work-records.jsonl row, a Git note, a
PR body section, or another named durable store. Raw session logs do not count.
/trace is a normal first-party skill under skills/, so bootstrap.sh
installs it system-wide through the existing cross-harness skill scan. No
special harness-native bridge is required.
/trace rejects obvious leaks but
does not summarize or sanitize a raw transcript for you.testing
Capture one compounding repo-technical learning while a solved problem is still fresh. Use when: after a bug fix, diagnosis, delivery, review, or incident reveals a reusable pattern worth adding to `docs/solutions/`. Trigger: /compound, /capture-learning, /learning.
testing
Route Misty Step factory application capabilities. Use when choosing, auditing, integrating, or operating Canary, Powder, Landmark, Aesthetic, or Bitterblossom: production observability, incidents, health checks, error logging, backlog/work-card state, release intelligence, UI/UX system adoption, or supervised/unsupervised agent dispatch. Trigger: /factory-apps, /factory-stack.
testing
Prove a skill beats no-skill with a falsifiable A/B eval, or retire it. Design, generate, run, and maintain a skill-specific eval: name the one claim the skill must earn, run it skill-on vs raw same-model, grade blind with objective checks first, return a keep/adapt/cut verdict. Use when: "eval this skill", "does this skill help", "prove the skill beats no skill", "write an eval for", "benchmark a skill", "is this skill worth it", "skill A/B", "skill regression test", "generate skill evals". Trigger: /skill-eval, /eval-skill, /prove-skill.
tools
> Template. Copy to `<target-repo>/.agents/skills/<repo>-<domain>/SKILL.md` > and fill every bracketed placeholder from the live target repo. Delete this > line and every other `> ` guidance line before committing. See > `../../references/repo-local-skill-generation.md` for the full process. --- name: <repo>-<domain> description: | [One paragraph: what this skill verifies/runs/operates for <repo>, stated in terms of the repo's real shape (service/CLI/library/etc.), not generic process. En