The Agent Skills Directory

[COMMAND_EXECUTION]: The skill defines a shell command workflow (opentaint scan) that directly interpolates user-provided inputs like <model-dir>, <report.sarif>, and <rules-dir>. Without explicit sanitization or validation of these inputs, an attacker could supply malicious values containing shell metacharacters (e.g., ;, &&, or backticks) to execute arbitrary commands on the host system.
[PROMPT_INJECTION]: The skill facilitates indirect prompt injection by ingesting untrusted data from project sources and rule directories.
Ingestion points: Data is read from user-specified paths for <project-src>, <rules-dir>, <config-dir>, and <approx-dir> in SKILL.md.
Boundary markers: Absent. The instructions do not provide delimiters or warnings to the agent to ignore instructions embedded within the ingested files.
Capability inventory: The skill executes the opentaint CLI tool, which interacts with the file system to read source code and write scan reports.
Sanitization: Absent. There are no instructions to validate, filter, or sanitize the content of the project being scanned or the rules being applied.

run-scan