feat(0.2): detector capability metadata + missing-input diagnostics (Tracks 9.1/9.3)

[pmclSF]

Summary

Two paired Track 9 deliverables that lift detector self-description
from "what does it emit" to "what does it consume" — and surface
input gaps as visible diagnostics instead of silent zero-output.

• Track 9.1 — Capability metadata. Adds 5 new DetectorMeta
  fields: RequiresRuntime, RequiresBaseline, RequiresEvalArtifact,
  ContextAware, Experimental. All zero-default; existing detectors
  unchanged.
• Track 9.3 — Missing-input diagnostics. New safeDetectChecked
  emits a single SignalDetectorMissingInput marker per affected
  detector when its declared inputs aren't satisfied. Skips
  invocation; no panic; explanation names the exact flags to add.

What changed

New code:

• DetectorMeta.{RequiresRuntime,RequiresBaseline,RequiresEvalArtifact,ContextAware,Experimental}
  — 5 new fields, all zero-default
• safeDetectChecked(reg, snap, fn) — canonical invocation path,
  composes input check with existing safeDetect panic recovery
• missingInputs(meta, snap) — returns user-facing flag names
  for each missing input
• joinInputNames — Oxford-comma string join (with tests
  covering 0/1/2/3+ cases)

New signal type:

• SignalDetectorMissingInput registered in signal_types.go,
  signal_catalog.go, manifest.go

Wiring:

• All 6 safeDetect(reg, ...) call sites in detector_registry.go
  (Run + RunWithGraph Phase 1/2/3 paths) routed through
  safeDetectChecked(reg, snap, ...)

New doc: docs/rules/engine/detector-missing-input.md (regenerated)

Test plan

• 7 new tests in missing_input_test.go — happy path,
  per-input missing diagnostics, multi-missing aggregation,
  join formatting
• go test ./... — full suite green
• make docs-verify — generated docs current

Plan tracker

Closes Track 9.1 + 9.3. Track 9 remaining: 9.2 (panic recovery
completion), 9.5 (pipeline architectural separation), 9.6
(registry refactor), 9.7 (truth-verify). All explicitly
post-0.2.0-blocking.

🤖 Generated with Claude Code

[github-actions]

[RISK] Terrain — Merge with caution

High-severity gaps found in changed code.

Metric	Value
Changed files	7 (4 source · 1 test)
Impacted units	32
Protection gaps	11
Tests selected	8 of 773 (1% of suite)

---

Coverage gaps in changed code

• internal/models/signal_catalog.go [MED] — Exported function KnownSignalTypes has no observed test coverage.
  → Add unit tests for exported function KnownSignalTypes — this is public API surface.
• internal/models/signal_catalog.go [MED] — Exported function SignalCatalogEntry has no observed test coverage.
  → Add unit tests for exported function SignalCatalogEntry — this is public API surface.
• internal/models/signal_catalog.go [MED] — Exported function SignalSource has no observed test coverage.
  → Add unit tests for exported function SignalSource — this is public API surface.
• internal/signals/detector_registry.go [MED] — Exported function All has no observed test coverage.
  → Add unit tests for exported function All — this is public API surface.
• internal/signals/detector_registry.go [MED] — Exported function ByDomain has no observed test coverage.
  → Add unit tests for exported function ByDomain — this is public API surface.
• internal/signals/detector_registry.go [MED] — Exported function Detectors has no observed test coverage.
  → Add unit tests for exported function Detectors — this is public API surface.
• internal/signals/detector_registry.go [MED] — Exported function Len has no observed test coverage.
  → Add unit tests for exported function Len — this is public API surface.
• internal/signals/detector_registry.go [MED] — Exported function Register has no observed test coverage.
  → Add unit tests for exported function Register — this is public API surface.
• internal/signals/detector_registry.go [MED] — Exported function Run has no observed test coverage.
  → Add unit tests for exported function Run — this is public API surface.
• internal/signals/detector_registry.go [MED] — Exported function RunDomain has no observed test coverage.
  → Add unit tests for exported function RunDomain — this is public API surface.
• ...and 1 more (1 medium)

8 pre-existing issues on changed files

• docs/signals/manifest.json [MED] — [aiModelDeprecationRisk] model tag gpt-3.5-turbo is a moving alias; pin a dated variant
• docs/signals/manifest.json [MED] — [aiModelDeprecationRisk] model tag gpt-4 resolves to whatever the provider currently maps it to; pin a dated variant (e.g. gpt-4-0613)
• internal/signals/manifest.go [MED] — [aiModelDeprecationRisk] model tag gpt-3.5-turbo is a moving alias; pin a dated variant
• internal/signals/manifest.go [MED] — [aiModelDeprecationRisk] model tag gpt-4 resolves to whatever the provider currently maps it to; pin a dated variant (e.g. gpt-4-0613)
• internal/models/signal_catalog.go [HIGH] — [blastRadiusHotspot] Changes to this file propagate to 2142 tests (1510 direct, 632 indirect). High blast radius increases regression risk.
• ...and 3 more

---

Recommended tests

8 test(s) with exact coverage of 21 impacted unit(s). 11 impacted unit(s) have no covering tests in the selected set.

Test	Confidence	Why
internal/engine/registry_error_test.go	exact	exact coverage of DetectorMeta, DetectorRegistration, Domain + 1 more
internal/engine/registry_test.go	exact	exact coverage of Domain, EvidenceType
internal/measurement/measurement_test.go	exact	exact coverage of NewRegistry
internal/signals/detector_registry_test.go	exact	exact coverage of IsKnownSignalType
internal/signals/manifest_rule_docs_test.go	exact	exact coverage of AllSignalTypes, Manifest, SignalStatus + 8 more
internal/signals/manifest_test.go	exact	exact coverage of SignalCatalog
internal/signals/missing_input_test.go	exact	exact coverage of DetectorMeta, DetectorRegistration, Domain
internal/structural/structural_test.go	exact	exact coverage of EvidenceType

---

AI Risk Review

Scenarios: 0 of 16 selected

2 advisory findings

• docs/signals/manifest.json:1087 — Model tag is sunset or floats — the next API call could break or silently re-resolve.
  → Pin to a dated model variant (e.g. gpt-4-0613) or upgrade to a current tier.
• internal/signals/manifest.go:841 — Model tag is sunset or floats — the next API call could break or silently re-resolve.
  → Pin to a dated model variant (e.g. gpt-4-0613) or upgrade to a current tier.

---

Owners: PMCLSF

Limitations

• No coverage artifacts provided; protection gaps reflect missing data, not measured absence. Provide --coverage to improve accuracy.
• Mixed test cultures reduce cross-framework optimization confidence. Consider standardizing on fewer frameworks.

_{Generated by Terrain · terrain pr --json for machine-readable output}

Targeted Test Results

Terrain selected 8 test(s) instead of the full suite.

• Go tests: passed

[github-actions]

Terrain AI Risk Review

Metric	Value
AI surfaces	13
Eval scenarios	16
Impacted scenarios	0
Uncovered surfaces	13

Decision: PASS — AI surfaces are covered.