setup-build-tools
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a shell script located at
$CLAUDE_PROJECT_DIR/.claude/hooks/install-build-tools.sh. While intended to install legitimate system dependencies likep7zip-full,wget, andnodejs, this pattern executes local code from a hidden directory. - [INDIRECT_PROMPT_INJECTION]: The file
SKILL.mdincludes the$ARGUMENTSplaceholder under the "Optional Guidance" section. - Ingestion points: The
$ARGUMENTSvariable inSKILL.mdallows external data to be interpolated into the prompt. - Boundary markers: There are no delimiters or instructions provided to the agent to treat the content of
$ARGUMENTSas untrusted data. - Capability inventory: The skill has the capability to execute shell commands (
bash). - Sanitization: There is no evidence of sanitization or filtering of the input provided via
$ARGUMENTSbefore it is presented to the agent.
Audit Metadata