The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill instructs the agent to execute echo "${TUZI_API_KEY:-not_set}" in SKILL.md, which explicitly prints sensitive credentials to the standard output and history.
[CREDENTIALS_UNSAFE]: The skill accesses sensitive environment configuration files located at .tuzi-skills/.env and $HOME/.tuzi-skills/.env in SKILL.md to check for and potentially read API keys.
[COMMAND_EXECUTION]: The skill executes shell commands using npx -y bun in SKILL.md to run a TypeScript file from a directory determined by another skill's metadata, introducing a dependency on external code paths.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user content and uses it to generate prompts and video generation arguments without sanitization.
Ingestion points: User-provided topic descriptions, articles, and scripts processed in Step 1 of SKILL.md.
Boundary markers: No delimiters or warnings are used to separate user-provided content from agent instructions.
Capability inventory: The skill performs shell command execution via npx and writes generated content to the prompts/ directory as described in SKILL.md.
Sanitization: There is no evidence of filtering, escaping, or validation of user-provided content before it is processed.

tuzi-short-video