V1.1

Affected components

• Repository documentation and cross-platform checkout rules.
• HookPrompt, Agent Teams Playbook, Find Skill, GoalPro, Kim Decision, and Semgrep Skill packaging metadata.

User-visible changes

• Rewrote the Chinese and English homepages to state clearly that Kim Service is a collection personally built, adapted, open-sourced, and maintained by Lao Jin (KimYx0207).
• Simplified the public README around what each project does, how to use it, where to get updates, and how to contact or support the maintainer.
• Moved catalog, content-hash, validation-gate, and release-protocol details out of the user-facing README and into docs/maintenance.md.
• Added repository-wide LF checkout rules so component snapshots verify consistently on Windows, macOS, and Linux.

Breaking changes and migration

• No user-facing breaking changes.
• V1.0 remains immutable, but fresh clones could report component hash drift when Git converted line endings on Windows. Use V1.1 or later for a reproducible checkout and verification result.

Verification

• Root repository, shared component runner, and release-contract checks passed after line-ending normalization.
• All six affected component hashes were recalculated from LF-normalized files; the three canonical direct-sync components remained byte-identical.
• A default Windows checkout from the remote V1.1 tag passed the repository and component gates.
• README layout checks passed with the contact banner centered at 720px and both 260px payment codes centered in one row.

Source revisions

• Component source revisions are unchanged from V1.0; this release changes repository documentation, checkout normalization, and packaging hashes only.

V1.0

Affected components

• Hook: HookPrompt.
• Skills: Agent Teams Playbook, Memory 3-Layer, Find Skill, GoalPro, Kim Decision, Meta Skill Creator, Semgrep Skill, and Xiaohongshu Skill.

User-visible changes

• Launched Kim Service as the consolidated public collection for one Hook and eight self-contained Skills.
• Included Agent Teams Playbook as a self-contained Skill under skills/agent-teams-playbook.
• Made every Skill directory independently understandable and installable with its own SKILL.md, README, LICENSE, CHANGELOG, NOTICE, and required runtime files.
• Standardized component verification behind one catalog-driven runner and removed component-specific validation exceptions from the root README.
• Standardized public README support visuals through one catalog-projected layout contract: the 720px contact banner is centered independently, while the two 260px payment codes share one centered row with centered cells.
• Replaced the residual fixed candidate-file checklist in Meta Skill Creator with conditional core / conditional / release rules.
• Renamed the Claude-specific memory component to platform-neutral memory-3layer, with one shared core, Claude Code and Codex Hook adapters, an explicit manual route, and a non-destructive legacy-data migration.
• Added direct canonical-to-Kim-Service exact projection for Meta Skill Creator, Memory 3-Layer, and Xiaohongshu Skill, with no persistent SKILL-* intermediary repository and no wrapper-composition layer.
• Added catalog-pinned component hashes, public-boundary checks, secret and machine-path scanning, nested-Git and runtime-projection checks, and protected QR asset verification.
• Retained the required contact QR, WeChat Pay QR, and Alipay QR assets.
• Established the exact repository release contract: VERSION, annotated tag, GitHub Release tag, and GitHub Release title use the same case-sensitive V<major>.<minor> string.

Breaking changes and migration

• The public repository identity changes from KimYx0207/agent-teams-playbook to KimYx0207/Kim_Service. Existing clones should update origin to the new repository URL.
• GitHub repository-rename redirects preserve the former KimYx0207/agent-teams-playbook web and Git URLs; the old repository name must not be reused because doing so would disable those redirects.
• The historical lowercase v4.8.0 tag and Release remain part of the preserved Agent Teams Playbook history. Kim Service collection releases use a separate uppercase V<major>.<minor> namespace beginning with V1.0.
• Users of former component repositories should use the corresponding hooks/<slug> or skills/<slug> path in Kim Service for current consolidated releases.
• Persistent SKILL-* intermediary directories and wrapper-based public export are no longer part of the Meta Skill Creator or Xiaohongshu Skill publication workflow.

Verification

• node scripts/check-repository.mjs passed for 9 components, 242 repository files, and 3 protected QR assets.
• Canonical/runtime/Kim Service direct-sync and full relative-path/SHA-256 gates passed for Meta Skill Creator, Memory 3-Layer, and Xiaohongshu Skill.
• Memory 3-Layer package validation and 27 installer, migration, recording, poisoning-resistance, and loader-preservation tests passed; real Claude Code and Codex Hook smoke remains a separate runtime proof layer.
• All catalog-declared component checks passed through the shared node scripts/check-components.mjs runner; components without an independent command remained covered by the same repository root gate.
• The exact V1.0 release-contract regression passed, including rejection of legacy or malformed collection version forms and empty CHANGELOG sections.
• The release gate accepts the manifest's owner/repository identity while still requiring origin to resolve to the matching GitHub repository URL.
• README layout regressions passed for incorrect banner width, missing centered containers, uncentered payment tables or cells, and payment codes split across rows.
• node scripts/check-repository.mjs --release is required to pass on the clean main release commit before tag creation.
• Remote branch, tag, Release metadata, and fresh-tag-clone checks are required after publication and are reported separately from local readiness.
• Superseded by V1.1 for fresh-clone verification because the original catalog hashes of six imported components reflected a CRLF working tree rather than the LF Git blobs.

Source revisions

• HookPrompt: a4c1faac0cc79860308f5553e3be0b0ac32415bb.
• Agent Teams Playbook history base: 753ff43bd9b1f9aee4d184c4f21e7f494af5a79f.
• Memory 3-Layer legacy source base: 1d60800c1a34dfe83d8e2b102b24c7d57d87ca53.
• Memory 3-Layer canonical direct-sync tree: 2592d1f68b3a005355148b5cdccceedeb74d7755e037970087639602537832fb.
• Find Skill: cf7635e3755c47b472bfb6dfd854680b5662ee26.
• GoalPro: 39adc8db765e0e4ad4df8d4ce02e7059fed69f26.
• Kim Decision: fbbe41cb6155ffd605c65b5af3f876ec25cfc0ea.
• Meta Skill Creator canonical direct-sync tree: 294245547c7ce33062926329e361c08dee4218bf0b33ada25d1c66a6cd39319b.
• Semgrep Skill: eb6dd5127f5dedc325b9364edf71a2034e5e35b1.
• Xiaohongshu Skill canonical direct-sync tree: 9464b5dab5222a671b546a5dc1b3e73f45f19f53e4bb54fe3b3b2401a2b537ca.

v4.8.0

Solved Problem

The v4.7 playbook forced an external Skill-search ritual and could label a successful native Agent dispatch as fallback when an optional Skill was not installed. That made healthy Codex and Claude Code execution look degraded.

Changes

• Stop capability discovery on the first qualified local Agent, Skill, Tool, Command, MCP provider, or capability-index match.
• Run external find-skills discovery only after a real local capability gap is proven.
• Use Codex's current top-level spawn_agent(task_name, message, fork_turns) contract without legacy typed or namespaced fields.
• Keep Claude Code on its native Agent / Task contract, require the Agent prompt, and use Team surfaces only when the host exposes them.
• Add synchronized Claude Code and Codex runtime packages, release metadata, regression tests, and repository ignore rules for secrets and local runtime state.

Verification

• node --test tests/runtime-contracts.test.mjs — 6/6 passed.
• bash -n scripts/install.sh — passed.
• Live Claude Code 2.1.196 regression — one Skill call and one Agent call with required prompt, no retry.
• Codex runtime evidence — three successful native top-level Agent spawns.