The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill provides instructions to download and execute installation scripts for well-known developer tools.
Evidence: curl https://raw.githubusercontent.com/nektos/act/master/install.sh | sudo bash and bash <(curl https://raw.githubusercontent.com/rhysd/actionlint/main/scripts/download-actionlint.bash).
These scripts originate from the official repositories of established community tools (act and actionlint) used for GitHub Actions validation.
[PRIVILEGE_ESCALATION]: The recommended installation procedure for the act tool requires sudo privileges to execute the downloaded shell script.
[COMMAND_EXECUTION]: The skill generates a local shell script (.github/scripts/validate-workflows.sh) that executes standard linting and validation commands (actionlint, yamllint, grep).
[INDIRECT_PROMPT_INJECTION]: The skill identifies a surface for indirect prompt injection by ingesting untrusted data from external sources.
Ingestion points: The skill reads content from .github/workflows/*.yml files and processes arbitrary error messages provided by the user during debugging sessions.
Boundary markers: The skill does not define specific delimiters or instructions to ignore embedded commands within the ingested workflow data.
Capability inventory: The skill has the capability to write files and execute shell commands through the generated validation scripts.
Sanitization: There is no evidence of sanitization or strict schema validation for the ingested YAML content or error logs.

github-actions-testing