From mutagen
Explicit-only skill. Read-only report on the mutagen workflow — upstream document status, April's Readiness Brief, Shredder's Validation Report, pipeline mode, slice queue progress, harness queue validation, active slice, heartbeat telemetry, gate telemetry, open escalations, recent reviews. Invoke only when the user explicitly says $mutagen-status.
How this skill is triggered — by the user, by Claude, or both
Slash command
/mutagen:mutagen-statusThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Strictly read-only. Collect state, report concisely, recommend the next
Strictly read-only. Collect state, report concisely, recommend the next action.
bash "$MUTAGEN_ROOT/scripts/status.sh" --format markdown.docs/ conventions and repo root) and read its
status line. Report path, status (Draft / In Review / Approved / Accepted
/ Missing), Last reviewed date if present, count of <TBD> markers..mutagen/state/readiness-brief.json
if present. Report date, per-document status, Shredder readiness
projection (green / yellow / red), recommendation. If present, trust it
over re-computing from raw docs. If absent, fall back to step 1..mutagen/state/validation-report.json
if present. Report date, bundle_ready, summarise readiness_issues
and validation_findings. Flag as stale if the bundle has been edited
since..claude/workflow.json. Report mode
(full / lightweight), review.max_retries, heartbeat thresholds.
Absent = default full, default retries = 2.slices/queue.json; fall back to
slices/slicemap.md when JSON is missing, and only then to legacy
slices/queue.md if needed. Report:
pending, in_progress, completed, refused,
escalated, blocked_retry.attempts, review_required (in lightweight mode)..mutagen/state/queue-validation.json if
present. Treat it as the harness verdict on whether slices/queue.json
is executable. Report:
ok, error_count, warning_count.level, code, slice_id when present,
and message.queue_contract_hash metadata, recompute the current queue-contract
hash and compare that instead of file modified times. Runtime-mutated
fields such as status, attempts, micro_corrections_used,
verdicts, completed_at, and escalation_reason do not make
the validator stale. Fall back to modified-time comparison only for
legacy reports without hash metadata.slices/queue.json is missing but the validator report exists,
flag it as orphaned..mutagen/state/active-slice.json if present.
Report slice ID, stage, active_agent, host, attempts, and any
degraded_capabilities. If the file exists outside a
$mutagen-execute-next run, flag it.bash "$MUTAGEN_ROOT/scripts/heartbeat.sh" 300. Report the JSON
(total, window_calls, bytes_last_window, last_run_length). If
last_run_length >= LOOP_THRESHOLD (default 5), flag a likely tool-call
loop.slices/queue.json, report Bishop and Tiger
Claw verdict counts across the last 10 completed slices.slices/queue.json, list slices whose
status is escalated, refused, or blocked_retry with their
escalation_reason.reviews/ and list the last three.mutagen workflow status — {YYYY-MM-DD HH:MM}
Upstream documents:
PRD {status} ({path}) · TBDs: N · last reviewed: YYYY-MM-DD
ADR {aggregate status} · {count} accepted / {count} draft
DDD {status} ({path}) · TBDs: N · last reviewed: YYYY-MM-DD
ISC {status} ({path}) · TBDs: N · last reviewed: YYYY-MM-DD
DSD {status} ({path}) · TBDs: N · last reviewed: YYYY-MM-DD
April Readiness Brief ({date} · {mode}):
Recommendation: {...}
Shredder readiness: PRD 🟢 · ADR 🟢 · DDD 🟡 · ISC 🔴 · DSD 🟢
Cross-doc issues: N (or "none noted")
(or: no readiness brief on file — run $mutagen-elicit)
Shredder Validation Report ({date}):
Bundle ready: true | false
Readiness issues: N · Validation findings: N
(or: no validation report on file — run $mutagen-slice)
Pipeline mode: full | lightweight · max retries: N
Slice queue (slices/queue.json):
Total: N · pending: N · in_progress: N · completed: N · blocked_retry: N · refused: N · escalated: N
By layer: L1: N · L2: N · L3: N · L4: N · L5: N · L6: N
Next up: {slice-id} · {author_agent} · L{n} · attempts {k} · "{objective}"
Queue validation ({path}):
Executable: true | false (flag stale / orphaned when applicable)
Errors: N · warnings: N
Issues: none noted
- [error|warning] {code} · {slice-id?}: {message}
(or: no queue validation report on file — run $mutagen-slice)
Active slice:
{slice-id} · stage: {stage} · agent: {active_agent} · host: {host} · attempts {k}
Degraded host features: {comma-separated list} (omit when empty)
Heartbeat (last 5 min): calls {N} · bytes {N} · last-run {tool}×{len}
(⚠ tool-call loop detected) | (⚠ stalled) | (nominal)
(or: no active slice)
Latest scope violation:
{slice-id} · stage: {stage} · agent: {active_agent} · class: {class}
path: {path}
recorded: {ts}
artifact: .mutagen/state/scope-violation.json
(or: none)
Recent gate telemetry (last N completed slices):
Bishop: clean M · advisory N · block 0 · skipped K
Tiger Claw: clean M · gap N · defect 0 · skipped K
Open escalations: {count}
- {slice-id} [{status}]: {escalation_reason}
Recent reviews (last 3):
- {slice-id}: 🟢/🟡/🔴 — reviews/{slice-id}.md
Next actions:
- {e.g. "PRD still Draft — run $mutagen-elicit"}
- {e.g. "Queue validation stale — re-run $mutagen-slice before dispatch"}
- {e.g. "Queue invalid — fix Shredder output before $mutagen-execute-next"}
- {e.g. "Queue clear — design bundle may need a new elicitation round"}
- {e.g. "Escalation pending on L3-Auth-0001 — awaiting user decision"}
.mutagen/state/active-slice.json
rather than modifying it.$mutagen-elicit.$mutagen-slice..mutagen/state/queue-validation.json is missing, stale, orphaned, or
reports ok: false, recommend $mutagen-slice instead of
$mutagen-execute-next.ok: true, recommend $mutagen-execute-next..mutagen/state/scope-violation.json exists, surface it even when the
queue already shows the slice escalated — it is the canonical detail
payload for the latest Traag DENY.degraded_capabilities, surface them plainly.
"Serial only" and "advisory scope" are runtime facts, not folklore.last_run_length, recommend investigating.npx claudepluginhub chkdsklabs/mutagen --plugin mutagenGuides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.