The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local shell script check-update.sh during the preamble stage using the current script's directory as a base path.
[REMOTE_CODE_EXECUTION]: The skill attempts to locate and execute a Python script named anysearch_cli.py by scanning multiple sensitive hidden directories in the user's home folder, including ~/.claude/skills/, ~/.cursor/skills/, and others. This involves executing code found at these dynamically computed paths.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) as it ingests untrusted product data and names from the user and interpolates them into analysis reports and logic.
Ingestion points: Product names and metrics collected via AskUserQuestion in Steps 1 and 2.
Boundary markers: No specific delimiters or safety instructions are used when handling the collected product data.
Capability inventory: The skill has access to Bash and Write tools, allowing it to execute commands and modify the file system based on the analysis.
Sanitization: There is no evidence of validation or sanitization for the user-provided product strings before they are used in the document generation process.
[COMMAND_EXECUTION]: The skill includes instructions to execute shell commands to create directories and check for the existence of strategic documents in the project path.

pm-portfolio