The Agent Skills Directory

[COMMAND_EXECUTION]: The skill performs several high-privilege administrative actions on the host repository using the GitHub CLI (gh).
It executes gh repo edit to change repository visibility to public, which is a high-impact operation that exposes all historical commit data.
It modifies repository settings to enable Private Vulnerability Reporting and configures branch/tag rulesets via gh api and gh repo edit.
It uses gh skill, a GitHub CLI extension, to validate and publish repository metadata.
[EXTERNAL_DOWNLOADS]: The skill uses npx to download and execute the skills package from the npm registry.
The commands npx skills add . --list and npx skills add <owner>/<repo> download and execute code from an external registry during the publication and verification phases.
[DATA_EXFILTRATION]: The skill facilitates the exposure and transmission of repository data.
The 'visibility flip' operation permanently exposes previously private source code and metadata to the public internet.
The verification step performs a 'seed install' which transmits repository metadata to skills.sh as telemetry to list the repo in the public catalogue.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and acts upon content from the repository files, creating a surface for indirect instructions to influence high-privilege actions.
Ingestion points: The agent reads and parses skills.sh.json, markdown files in the skills/ directory, the README, and the CHANGELOG.
Boundary markers: The instructions do not define clear delimiters or 'ignore' instructions for the content of these files, which are processed during pre-flight checks.
Capability inventory: The skill has access to repository configuration commands (gh repo edit), network access (npx), and script execution (python3, npx).
Sanitization: There is no explicit validation or sanitization of the content of the repository files before they are used to determine deployment readiness or grouping logic.

publish-repo