[abruption]

Closes #32

Summary

• Add the explicitly reviewed public runtime 0.3.4 legacy/successor matrix row and fixed-source publication evidence.
• Update the TS setup and canonical instructions to recognize reviewed runtimes 0.3.3 and 0.3.4. Retain the historical 0.3.3 pure-validator rejection of a supplied 0.3.4 identity and explain why the actual published 0.3.4 guard differs.
• Explain Windows PowerShell SSH native exit-code forwarding, retaining complete validated submission evidence after abnormal transport exits, and uncertain partial/malformed output without automatic resend.
• Preserve exact-version endpoints, existing package-manager ownership, separate skill updates and submission != consumption/ACK.
• Keep TS frontmatter byte-identical: skill 0.2.0 / minimum 0.1.0 / full 0.3.2 / typescript / probe-help. Python canonical bytes and README locales are unchanged.
• Keep this work separate from Python PR docs: 외부 메시지와 Reply-To 신뢰 경계 명확화 #30 and future TS 0.4.0 planning issue docs: Prepare TypeScript 0.4.0 handoff guidance and staged compatibility rollout #31.

Evidence

• Public 0.3.4 release, npm registry latest=0.3.4, and exact source afed4a5a8945012faf1b23d52cffecb70470a5b3 were checked.
• The 0.3.3..0.3.4 guard source change explicitly adds only the reviewed 0.3.4 row; it does not introduce wildcard acceptance. Public CLI/SSH response source was also read.
• Artifact hashes/provenance in the compatibility document are attributed to the respective publication records. This documentation task did not independently install the npm runtime artifact or run live SSH/agent ACK qualification.

Validation

• node scripts/validate-skill.mjs: PASS, including metadata and local-link checks.
• skill-creator quick_validate.py session-peer-ts: PASS using a temporary venv with PyYAML 6.0.3.
• node scripts/test-ts-skill.mjs: PASS, isolated project/global installation, source replacement, listing and removal; Python fixture preserved.
• git diff --check: PASS.
• Exact comparisons against origin/main: TS frontmatter and Python canonical skill unchanged; only the three listed TS files changed.
• Work split between a bounded Codex subagent and primary Codex; the primary reviewed the diff and finite-matrix source change. No external Claude Code/Antigravity CLI was used for this release-guidance task.
• Current head cfbe95f5579bcdd38f15074c2263bcf0b58ec136: CI 2/2 PASS (skill-package / validate, GitGuardian). Automatic code/security review completed on the same head with no reported findings or inline comments; previous-head checks are not reused.

Scope and release boundary

Documentation only: session-peer-ts/SKILL.md, docs/typescript.md, docs/typescript-compatibility.md.
No merge, tag, release, npm publication, user-installed skill/runtime update, Python snapshot edit or operations change.
Existing immutable session-peer-ts-v0.2.0 artifacts and adoption pins remain unchanged. These edited instructions are a documentation candidate, not replacement release bytes or approval to finalize/publish a new skill/profile version.

Branch: docs/ts-0-3-4-guidance
Head: cfbe95f5579bcdd38f15074c2263bcf0b58ec136
Candidate TS SKILL SHA-256: 7cf706a0ec59da4f6c91054aeb599efd3be2aa90073dd1e91804aaca83e71de5

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-12T05:33:02.572649Z	cfbe95f	PR opened
🔒 Security Review	✅ Completed 2026-10-12T05:33:06.602465Z	cfbe95f	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.