Historical field note

This record preserves prior investigation evidence for a single dream finding. It is not a current operating contract; use the numbered book and current runbooks for instructions.

Investigation: dream finding `dreamrepair:ffbc63f8695e9316b064bb1f6d3566cb` (skill)¶

Task: Assemble the complete evidence record for a low-confidence dream-cycle failure_pattern finding so its actionability can be judged from the attached evidence alone. Review-only — no skills, tools, or source were changed. Finding: kind failure_pattern, scope project, project mac, fingerprint dreamrepair:ffbc63f8695e9316b064bb1f6d3566cb, confidence low (confidence_score = 0.35), evidence_count = 1, affected label skill (generic placeholder — no concrete skill/tool/provider/repo_area name). Classifier signal that fired: \bskill[s]?\b → area type skill (src/mac/dream_cycle_classifier.py:103). Sole supporting record: mem_88ebe4bba2924b97b1843560a58884ee (record_type deployment_learning:mac), from repair task task_repair_a2b91fa4f146c4d1f879485c. Origin dream memory: mem_75cf619496f94677aa31b0ad45d3adb2, emitted by nap run nap_15e3fb5ce5cf4ae0a88b464d22939995. Investigated by: fleet worker (investigation node; no production code, test, skill, or deploy edits).

Verdict: NOT ACTIONABLE — single-record, info-only inventory signal, not a defect¶

The finding rests on exactly one supporting record whose own claim is that skill:<name> dream candidates are info-only inventory records (severity="info"), not failure evidence. There is no failing test, stack trace, error signature, reproduction, or named offending skill/tool/provider behind it. The 0.35 score is the classifier's deterministic single-record structural floor (an evidence-volume signal), and the skill label is a bare word-boundary token match on \bskill[s]?\b, not a concrete defect location. This adopts and independently re-verifies the same ground truth reached for the near-duplicate skill findings already closed NOT ACTIONABLE in docs/archive/field-notes/investigation-dreamrepair-c8dd8037-skill.md and docs/archive/field-notes/investigation-dreamrepair-5404b15-skill.md.

No files under src/, tests/, skills/, or deploy/ were modified by this investigation.

Provenance chain (as retrieved)¶

The evidence flows in one direction, origin → supporting record → finding, with no independent corroboration entering at any hop:

Origin dream memory mem_75cf619496f94677aa31b0ad45d3adb2 was emitted by nap run nap_15e3fb5ce5cf4ae0a88b464d22939995. It is the consolidation-time dream artifact that seeds the low-confidence failure_pattern candidate.
Supporting record mem_88ebe4bba2924b97b1843560a58884ee (deployment_learning:mac) was authored by repair task task_repair_a2b91fa4f146c4d1f879485c. It is the single evidence item the dream classifier attaches to the finding (evidence_count = 1, record_type_counts: {deployment_learning:mac: 1}).
Finding dreamrepair:ffbc63f8695e9316b064bb1f6d3566cb is the stable dedupe key repair_fingerprint() hashes from the candidate's {kind, scope, project, signature, summary, affected} (src/mac/dream_repair_tasks.py:175). It is a dedupe key for a dream candidate, not a pointer into any single skill module.

What the sole supporting record actually claims¶

mem_88ebe4bba2924b97b1843560a58884ee does not report a broken skill. Its substantive claim is that skill:<name> dream candidates are pure name inventory, not a failure signal — they are emitted at severity="info" for every skill name that merely appears in session text, and carry no failure excerpt or root signal. That claim is exactly the behavior the current tree encodes and can be corroborated read-only against source:

The scanner emits skill:<name> as a tool_or_skill_name candidate at severity="info", with excerpt "skill referenced: <name>", for every skill name extracted from session text (src/mac/dream_scanner.py:256). The tool variants (tool:<name>) are emitted the same way (src/mac/dream_scanner.py:239, src/mac/dream_scanner.py:247).
DREAM_INVENTORY_ONLY_KINDS = frozenset({"tool_or_skill_name"}) marks this kind as "pure name inventory rather than a failure signal" (src/mac/dream_repair_tasks.py:20).
_is_inventory_only_candidate() skips such candidates before the low + affected gate precisely because they are severity="info" with "nothing to act on" (src/mac/dream_repair_tasks.py:216), and the scan report records them as status="skipped", reason="inventory_only_kind" (src/mac/dream_repair_tasks.py:109).

So the record's content is a statement about classifier/scanner inventory semantics, not an observation of a skill malfunction. Treating it as fresh failure evidence corroborates only itself.

Classifier signals that fired¶

Area signal: \bskill[s]?\b → area type skill (src/mac/dream_cycle_classifier.py:103, within _SKILL_PATTERNS). This is a plain word-boundary match on the token skill/skills, matched against summary + observations + record_type_counts keys. It is a text match, not a named skill.
Confidence: confidence_score = 0.35. That is the deterministic low floor CONFIDENCE_THRESHOLDS["low"] = ("low", 0.35) (src/mac/dream_cycle_classifier.py:87), returned by _confidence_for(...) whenever evidence_count < 2 and no second distinct signal type or record type is present (src/mac/dream_cycle_classifier.py:233). With exactly one evidence record the finding is necessarily low; medium needs two, high needs three.
Affected label: skill is a generic area bucket, not a concrete skill. affected.tools, affected.providers, and affected.repo_areas are empty.

Together these explain the finding's kind, scope, confidence, and label entirely from the consolidation heuristics acting on one info-only inventory record. None of it signals that a skill module misbehaves.

Corroboration assessment: single record only¶

The finding has one supporting record and no independent corroboration:

Single record — evidence_count = 1; record_type_counts: {deployment_learning:mac: 1}. medium/high confidence is structurally unreachable without a second, independent record.
Self-referential / info-only — the lone record's own claim is that these skill:<name> candidates are info-only inventory, so it cannot attest to a new defect; it corroborates only the inventory semantics it describes.
No reproduction — no failing test, no reproducer, and no concrete defect location (the skill label is a placeholder, not a module/tool/provider).

Determination¶

Do not open a skill-behavior repair from this finding: it is not actionable as a skill-subsystem defect.
Treat the finding as low-confidence / unsubstantiated; it can be aged out or superseded once no additional independent evidence accumulates.
Reopen only if a named skill or tool acquires a reproducible failure signature (a real error/stack/failing test) with at least two independent, non-self-referential evidence records.

Assumptions recorded¶

The originating hub was not reachable from the investigation sandbox (mac admin memory search returned HTTP 403 Forbidden), so the three referenced memory records were not live-queried. Their identifiers, record types, provenance, and quoted claim are taken from the task specification and corroborated against the current source tree, which is the same method used by the sibling skill field notes cited above.

Investigation: dream finding dreamrepair:ffbc63f8695e9316b064bb1f6d3566cb (skill)¶