capture validates related_specs against the retired fn-N namespace (reFnID) instead of the live spc-N; adr-11 renamed fn to spec, fn- is a record-lint blocker, and 0 of 340 records populate the field, so it is unusable for its documented purpose

capture validates related_specs against the retired fn-N namespace (reFnID) instead of the live spc-N; adr-11 renamed fn to spec, fn- is a record-lint blocker, and 0 of 340 records populate the field, so it is unusable for its documented purpose

Evidence

• .abcd/work/issues/README.md:51 — related_specs documented as "list of fn-N ids".
• internal/core/capture/validate.go:132 — {"related_specs", reFnID, "fn-N"}; reFnID = ^fn-[0-9]+$ (capture.go:240). Nothing accepts spc-N for this field.
• adr-11 renamed fn→spec; live store is spc-N (internal/core/lint/schema.go:50). .abcd/record-lint.json bans \bfn- (blocker flow-next-spec-prefix), escaped here only because record-lint roots = [.abcd/development].
• reSpcID already exists and is used (workflow.go:222, --spec/BySpec). Tests pin the retired form: workflow_test.go:110,117 set RelatedSpecs: []string{"fn-12"}.
• 0 of 340 issue records populate related_specs — the field is unusable for its documented purpose.

Adversarial verdict

CONFIRMED — code fix substantive (low impact), doc fix nitpick (must bundle with code). Refuter corrected two false sub-claims in the original hunter note (reSpcID is used, not dead; tests do pin fn-N) — the core drift survives both. Fix: validate.go:132 reFnID→reSpcID, "fn-N"→"spc-N"; README:51 fn-N→spc-N; update workflow_test.go:110,117 to spc-12 (watched-fail: spc-12 rejected before, accepted after).