Proposal
Add a maintained coderabbit-hol-guard Agent Skill to this repository's canonical skills/ distribution. The skill would directly install and invoke HOL Guard on a supported local coding-agent harness before state-changing tool work, while keeping CodeRabbit review behavior and host-native permissions authoritative.
This follows the repository's issue-first policy for a new skill. If the scope is accepted, I can contribute the smallest repository-native implementation and run the required validation before opening a PR.
Intended maintained artifact
The accepted skill should use the real HOL Guard runtime rather than prompt-only checks, for example:
pipx install hol-guard
hol-guard detect --json
hol-guard install <detected-harness>
hol-guard run <detected-harness> --dry-run
hol-guard run <detected-harness>
hol-guard doctor <detected-harness> --json
For package verification, it can keep the separate scanner boundary explicit:
pipx install plugin-scanner
plugin-scanner verify .
A deny, review-required, error, malformed result, or unavailable Guard state should not be treated as permission to retry the same mutation through an unprotected agent session. The skill would also state clearly that HOL Guard protects the supported local agent harness; it does not run inside CodeRabbit or replace CodeRabbit's review/security controls.
Distribution fit
DISTRIBUTION_CHANNELS.md identifies skills/ plus npx skills add coderabbitai/skills as the live canonical multi-agent distribution for 35+ compatible agents. That is the target surface for this proposal; I am not asking for a generic framework hook or an unrelated marketplace listing.
Scope / validation
- one focused Agent Skill plus only the repository-owned manifests/docs required by current packaging rules
- exact commands checked against the maintained HOL Guard skill/runtime
git diff --check and every repository validator applicable to the changed distribution surfaces- no claim that unavailable host-specific validators passed
- affiliation disclosure: I maintain HOL Guard / Hashgraph Online; AI assistance may be used to prepare the focused contribution, with the submitted behavior checked against current repository and HOL Guard documentation
HOL Guard: https://hol.org/guard
Source: https://github.com/hashgraph-online/hol-guard
Proposal
Add a maintained
coderabbit-hol-guardAgent Skill to this repository's canonicalskills/distribution. The skill would directly install and invoke HOL Guard on a supported local coding-agent harness before state-changing tool work, while keeping CodeRabbit review behavior and host-native permissions authoritative.This follows the repository's issue-first policy for a new skill. If the scope is accepted, I can contribute the smallest repository-native implementation and run the required validation before opening a PR.
Intended maintained artifact
The accepted skill should use the real HOL Guard runtime rather than prompt-only checks, for example:
For package verification, it can keep the separate scanner boundary explicit:
pipx install plugin-scanner plugin-scanner verify .A deny, review-required, error, malformed result, or unavailable Guard state should not be treated as permission to retry the same mutation through an unprotected agent session. The skill would also state clearly that HOL Guard protects the supported local agent harness; it does not run inside CodeRabbit or replace CodeRabbit's review/security controls.
Distribution fit
DISTRIBUTION_CHANNELS.mdidentifiesskills/plusnpx skills add coderabbitai/skillsas the live canonical multi-agent distribution for 35+ compatible agents. That is the target surface for this proposal; I am not asking for a generic framework hook or an unrelated marketplace listing.Scope / validation
git diff --checkand every repository validator applicable to the changed distribution surfacesHOL Guard: https://hol.org/guard
Source: https://github.com/hashgraph-online/hol-guard