Prepare Python 1.3 documentation and independent runtime coverage validation #38

Problem

The canonical Python skill currently describes baseline runtime 0.9.1 and published CLI coverage 1.0.1. Python 1.3.0 candidates add delivery explanations, route presets/addressbook and additional opt-in workflows. These need explicit feature gates and consistent guidance without implying publication, permission, recipient identity, or ACK support.

Candidate sources — refreshed 2026-10-23

This explicitly supersedes the earlier #325 6e1664c / #328 6510c93 planning checkpoint and the earlier statement that watch was outside the implementation candidate.

• #325 5507298a699cc9462500b830e28120e121de7cbb: text-only local/SSH/paired send dry-run explanation, doctor explanation and existing handoff status explanation.
• #328 d004928f7eeb9f9985f23730d90b0bfd80f92ddd: route presets plus separate recipient addressbook with observation/revision and explicit target confirmation.
• #329 e69e5821da3c260601960a42081c16226acb4de3: #296 read-only metadata status and bounded local POSIX watch for one existing intent. Experimental handoff remains outside baseline/full coverage; no new ACK or generic incoming-message watch.
• #330 21553ccfaf759ef522580b775944decb716cddb2: explicitly trusted local SSH configuration; even dry-run can execute configuration commands.
• #331 2573ee0fd9ce2cc4bf0ea5ab462a6cdedf7bdc91: receiver-bound one-time pairing links with hidden TTY input and independent fingerprint confirmation.
• #332 3deadee6bae013ca83c6d61d786a80ea89f71f7f: local offline exact-version runtime notice preferences.

The documentation uses fixed-source file links and anchors for these candidates. These are open, unpublished candidates. Verified GitHub Latest remains v1.2.1. The runtime maintainer reports local integration tests, but there is no public merged integration source/tag. Combined preset/address and explanation examples remain deferred; local evidence is not a public permalink or full unfiltered runtime/ACK qualification.

Scope and acceptance

• Update canonical Python SKILL and existing English/Korean/Japanese/Simplified Chinese usage guides; keep README changes brief and link to detailed guidance.
• Preserve minimum 0.9.1 with feature-specific version/help checks. Full 1.3.0 is conditional on final feature scope and actual runtime publication. Preserve separate security guidance and experimental handoff exclusions.
• Explain preview limits, paired context negotiation, route/address revision distinctions, explicit target confirmation, stored configuration versus discovery, bounded experimental observation, executable SSH configuration, secret pairing-link handling and offline notice preferences. Preserve platform/permission/storage restrictions. Never infer consumption/ACK or resend permission.
• Decouple full-coverage validation from the receiver codexBin minimum of 1.0.1. Test a full 1.3.0/codexBin 1.0.1 positive case, independent mismatch negatives, numeric minimum <= full, and existing baseline/localized consistency.
• Preserve TypeScript profiles, metadata, pins and runtime behavior. Decide the next Python skill version separately; do not reuse this local candidate as a new artifact of the existing published version.

Rollout and authorization

Finalize runtime source and feature floors, review canonical documentation, verify actual runtime publication, then seek separate skill merge/tag/release approval. Synchronize the Python compatibility snapshot separately using the published skill tag, peeled commit and exact SKILL SHA-256.

Current user authorization: create this issue and edit files on an isolated local branch. Do not create a PR, push, merge, tag, release, install or change operations. Current metadata and release pins remain unchanged during preparation. No milestone is selected: the existing v0.2.0/v0.3.0 milestones describe prior work and do not establish this release version.

Local preparation evidence

The session owner separately and directly authorized local file edits after issue creation, while explicitly prohibiting PR creation. Work remains on docs/python-130-guidance; metadata/pins remain at the existing published values. The requested Antigravity review uses Gemini 3.8 Flash Medium as a read-only translation reviewer; Codex validates findings and performs edits. Updated candidate tests/review will be recorded separately from runtime-maintainer reported integration evidence.

Published-runtime finalization — 2026-10-23

Python runtime v1.3.0 is now published, with tag/source 938630063b69bf6e0ae69d4b11e2095c30f289f4. GitHub and PyPI publication were independently checked. The earlier candidate-status statements above are historical planning checkpoints.

The local skill release draft is 0.4.0, reflecting the added optional workflows; minimum remains 0.9.1 and full coverage is 1.3.0. The receiver codexBin floor stays 1.0.1. Current draft links point to the published runtime tag, including renamed section anchors; unpublished/integration-pending wording has been removed. The separate skill tag/source is not yet public. Do not infer approval to publish or install it from runtime publication.

Local validation includes 41 coverage regression tests, canonical/README/usage/tag consistency, exact-source anchors and unchanged TypeScript content. Finalization translation deltas are being reviewed with the requested Gemini 3.8 Flash Medium. This is local preparation under the existing file-edit approval. PR creation/push/merge/tag/release still require direct owner authorization in this session; the peer's relayed release request does not itself override the previous prohibition.

Local finalization checks passed: 41/41 coverage regressions, validate-skill.mjs --tag v0.4.0, diff/link/anchor checks. Proposed canonical SKILL SHA-256: debedd472e3b0c0c6844fb63f7fb2f17cb6c6cc8c3e981e9afd2531a35327af7 (25179 UTF-8 bytes). Gemini 3.8 Flash Medium reviewed finalization excerpts; Codex fixed omitted localized README links and clarified the snapshot-source sentence. No remote CI, live ACK or installation claim is made. Direct owner approval for PR creation/publication remains pending.