afk-ask
Warn
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill defines shell command templates in
references/providers.md(e.g.,kiro-cli,opencode,codex,agy) that interpolate user-provided input via the{{PROMPT}}variable. This poses a risk of command injection if the user includes shell metacharacters in their request.\n- [INDIRECT_PROMPT_INJECTION]:\n - Ingestion points: Untrusted data enters the agent context through the user-provided
{{ARGUMENTS}}task inSKILL.mdand via the raw output from external AI CLI advisors.\n - Boundary markers: The skill lacks explicit boundary markers, structural delimiters, or 'ignore embedded instructions' warnings for user-provided data within the generated prompts.\n
- Capability inventory: The skill possesses shell execution capabilities (via templates in
references/providers.md) and file-writing capabilities (artifact creation inSKILL.md).\n - Sanitization: While the skill includes behavioral instructions in
SKILL.mdto treat external output as advisory and not a source of truth, there is no technical sanitization, validation, or escaping of the ingested content.\n- [DYNAMIC_EXECUTION]: Thereferences/providers.mdfile instructs the agent to inspect CLI help documentation at runtime and dynamically adapt command-line flags if defaults fail, which introduces unpredictability in the executed commands.
Audit Metadata