The Agent Skills Directory

[PRIVILEGE_ESCALATION]: The file scripts/generate_report.py contains a generate_script method that writes arbitrary content to a shell script and explicitly sets file permissions to 0o755 (executable) using file_path.chmod(0o755). This capability allows for the creation of executable binaries or scripts on the local filesystem.
[COMMAND_EXECUTION]: The scripts/generate_report.py utility can generate bash scripts using content provided via command-line arguments. This functionality, combined with the automatic application of execution permissions, provides a mechanism for arbitrary command execution if the input is manipulated.
[METADATA_POISONING]: The script scripts/analyze_headers.py contains a significant discrepancy between its documentation and its implementation. The file header and the README in scripts/ claim the script uses requests and beautifulsoup4 for security header analysis, but the provided source code only performs basic statistics on local files and directories.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external web domains provided by the user.
Ingestion points: External data enters the agent context via WebFetch and WebSearch when auditing website headers.
Boundary markers: The instructions lack explicit delimiters or warnings to treat fetched header content as untrusted data.
Capability inventory: The skill includes file-writing and execution permission capabilities through scripts/generate_report.py.
Sanitization: There is no evidence of sanitization or validation of the external content before it is used to generate reports or remediation configuration snippets.

analyzing-security-headers