chore(release): bump agent-runtime to 0.93.2

[drewstone]

What

• release the Codex local-execution repair as @tangle-network/agent-runtime@0.93.2
• update the canonical and generated version references

Included fix

• use codex exec for noninteractive worktree runs
• preserve composed system/task prompt and model selection
• forward AgentProfile.model.reasoningEffort to Codex, including xhigh

Proof

• pnpm test (1,500 passed, 2 skipped)
• pnpm typecheck
• pnpm lint (435 files)
• pnpm build
• pnpm verify:package
• pnpm docs:check

[tangletools]

✅ Auto-approved drewstone PR — 863fb91f

This PR was opened by the trusted drewstone account.
The full PR reviewer audit still runs separately and will publish findings if it detects issues.

_{tangletools · auto-approval · reason: drewstone_author · 2026-07-12T02:25:43Z}

[tangletools]

🟢 Value Audit — sound

Verdict	sound
Concerns	0 (none)
Heuristic	0.0s
Duplication	0.0s
Interrogation	64.1s (2 bridge agents)
Total	64.1s

💰 Value — sound

Standard release cut 0.93.1→0.93.2 for the already-merged Codex noninteractive-exec fix; bumps package.json plus the two version-pinned generated docs exactly as the freshness gate requires.

• What it does: Bumps @tangle-network/agent-runtime from 0.93.1 to 0.93.2 in package.json (line 4) and propagates that version into the two docs that carry the pin: docs/api/primitive-catalog.md (generated) and docs/canonical-api.md (hand-curated Version line). No code delta — the capability (codex exec instead of codex run, plus AgentProfile.model.reasoningEffort forwarding including xhigh) already landed in com
• Goals it achieves: Publishes the Codex local-execution repair as a consumable npm version so downstream packages can pin >=0.93.2. It also keeps docs/canonical-api.md and docs/api/primitive-catalog.md in sync with the source version, which is enforced — pnpm docs:check (package.json:103) runs docs:api regeneration + git diff --exit-code -- docs/api + docs:freshness (package.json:102), so a stale version pin
• Assessment: Textbook release bump in the established repo pattern (prior commits ad5093e '0.93.1', cfe92fc '0.91.0', 4229f80, e26c161 follow the identical chore(release): shape). Scope is correctly minimal: exactly the three files that must move together for a version bump — package.json + the two docs whose freshness gate hard-couples them to the source version. Nothing extraneous, no behavioral change smu
• Better / existing approach: none — this is the right approach. Verified the fix it releases (codex exec + reasoningArgs) is the single source of truth for Codex harness invocation at src/mcp/local-harness.ts (HARNESS_INVOCATIONS.codex + codexReasoningArgs + harnessInvocation), so there is no alternate release path or duplicate version surface to reuse; the bump-and-regen flow is the repo's canonical release mechanism.
• Model: opencode/zai-coding-plan/glm-5.2
• Bridge attempts: 2
• Bridge warning: opencode/kimi-for-coding/k2p7: bridge stream ended without value-audit content

🎯 Usefulness — sound

Routine patch release bump (0.93.1 → 0.93.2) shipping the already-merged Codex noninteractive-exec fix (#519); the version is consistent across all three references and the underlying fix is fully wired and reachable.

• Integration: The change under release is already merged (commit 07eed91, PR #519). This PR only bumps package.json:3 and regenerates the two version-citing docs (docs/canonical-api.md:5, docs/api/primitive-catalog.md:10) — all three now consistently read 0.93.2. The shipped capability (codex exec at src/mcp/local-harness.ts:76) is reachable via runLocalHarness/harnessInvocation from at least three live call
• Fit with existing patterns: Follows the established release pattern exactly — same shape as prior bumps (0.93.1 via #518, 0.91.0 via #510). Patch-level bump correctly reflects a fix-only change. The released fix uses the same HARNESS_INVOCATIONS table pattern shared by all three harnesses (claude/codex/opencode), adding the exec subcommand and reasoningArgs only to codex, consistent with the per-harness mapper design at sr
• Real-world viability: The version bump itself is a mechanical string change with zero runtime risk. The fix it releases handles the real execution path: stdin EOF closure (local-harness.ts:265), SIGTERM on timeout/abort (local-harness.ts:276,284), non-zero exit surfacing via exitCode, and the codexReasoningEffort map covers the full ReasoningEffort union including xhigh. The codex exec form is the documented noninter
• Model: opencode/zai-coding-plan/glm-5.2
• Bridge attempts: 1

No concerns — sound change, no better or existing approach found. ✅

---

What this audit checks

It judges the change on its merits — not whether it was tasked out in an issue. Unticketed, fast-moving work is fine; the question is whether the change is good and whether a better or existing approach should be used instead.

Pass	What it asks
Heuristic	Vague title? Whitespace-only or cruft-bearing diff? (content signals only)
Duplication	Do added function/class names already exist elsewhere in the repo?
Value Audit	What does it do? What goal does it achieve? Is it good? Better architecture or already-exists?
Usefulness Audit	Does it integrate and fit? Will it hold up in real use and actually get used?

Findings are concerns, not blocks — the human reviewer decides what to do with them.

_{value-audit · 20260712T022843Z}

[tangletools]

✅ No Blockers — 863fb91f

Review health 100/100 · Reviewer score 95/100 · Confidence 75/100 · 0 findings (none)

deepseek: Correctness 95 · Security 95 · Testing 95 · Architecture 95

Reviewer score is advisory once the run is complete and the verdict has no blockers.

Full multi-shot audit completed 3/3 planned shots over 3 changed files. Global verifier still owns final merge decision.

No findings.

---

_{tangletools · 2026-07-12T02:29:17Z · trace}

[tangletools]

✅ Clean — 863fb91f

Full multi-shot audit completed 3/3 planned shots over 3 changed files. Global verifier still owns final merge decision.

Full immutable report for this review: trace

Summary comment for this run: full summary

---

_{tangletools · 2026-07-12T02:29:17Z · immutable trace}