docs: record subagent outcomes for findings 2, 31, 38

Three findings from the simplification audit were independently
re-verified before execution, corrected, then implemented by
subagents:

- Finding 2 (check-executables-allow-sync): the audit's "drop it"
  option was found unsafe (ADR-0019 calls this failure mode silent,
  not "visible and recoverable" as claimed); shrunk instead of
  deleted, 231 -> 222 lines.
- Finding 31 (CONTEXT.md): "most terms unused by skills" was found
  overstated (13 of 28 are model-facing must-keeps); cut only the
  9 confirmed true orphans, 28 -> 19 terms. Also de-referenced one
  dangling pointer to a deleted term in the Flagged-ambiguities
  section.
- Finding 38 (pc-author/pc-run): line count was found overstated
  (598 actual vs. 689 claimed); trimmed the two generic reference
  files by 60 lines while preserving house-specific content.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01YR2CjVumUbEGWcMikcoXBD

CONTEXT.md

@@ -171,8 +171,8 @@ _Avoid_: ticket, card, task
 - Skills can answer to two names, bare (`gitea-prs`) and namespaced (`gitea:gitea-prs`), depending on
   whether a native install exists at user scope alongside the apm one (ADR-0018) — resolved: write
   the bare name, which is the only form `apm install` produces.
-- "context" means both the model's live token window (the **Preload tax** sense) and the bounded
-  domain this file describes — resolved: unqualified "context" in this repo means the token window.
+- "context" means both the model's live token window and the bounded domain this file describes —
+  resolved: unqualified "context" in this repo means the token window.
 - "audit" was used for both an author skill's inline closeout and `forge`'s independent
   clean-context recheck — resolved: these are two distinct layers, kept separate precisely because
   an audit running in the same context as the work it checks shares that work's blind spots.