The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill is explicitly designed to handle, capture, and store secrets. The instructions in SKILL.md direct the agent to read sensitive local files, including .env, .env.example, and GitHub workflow files (.github/workflows/*), to identify the specific variables and secrets required for the setup procedure.
[COMMAND_EXECUTION]: The generated bash script utilizes powerful CLI tools, specifically the GitHub CLI (gh), to perform actions like gh secret set and gh variable set. It also performs direct file system modifications to update environment files.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it ingests data from various repository files (such as README, docker-compose.yml, and framework configs) to determine the wizard's stages.
Ingestion points: Sensitive repo files like .env, .github/workflows/*, and documentation like README (specified in SKILL.md).
Boundary markers: No explicit delimiters or warnings to ignore instructions within the processed files are mentioned.
Capability inventory: The generated script can write to the filesystem (write_env), open browser URLs (open_url), and modify repository secrets via the GitHub CLI (set_secret).
Sanitization: No sanitization or validation of the content extracted from repo files is specified.
[DYNAMIC_EXECUTION]: The skill operates as a code generator, creating a new bash script by populating a provided template.sh with custom logic. This generated script is intended to be executed locally on the user's machine.
[PRIVILEGE_ESCALATION]: The instructions in SKILL.md explicitly direct the user or agent to grant execution permissions to the newly generated script using chmod +x <script>, which is a common step for dynamic script execution.

light-wizard