The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill promotes the installation of a global Node.js package 'task-master-ai' via npm. Installing packages from unverified sources is a risk as it grants execution persistence on the host system.
[COMMAND_EXECUTION]: Extensive use of the Bash tool to execute CLI operations such as 'task-master init', 'parse-prd', and 'analyze-complexity' which grants the tool significant control over the local project environment.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it ingests untrusted data from the '.taskmaster/docs/prd.md' file to generate task lists.
Ingestion points: The PRD file path '.taskmaster/docs/prd.md' is directly parsed by the CLI tool.
Boundary markers: No specific delimiters or safety instructions are provided to the agent to ignore potentially malicious content inside the PRD.
Capability inventory: The agent has access to 'Bash', 'Read', and 'Write' tools, allowing it to execute commands or modify files based on instructions parsed from the PRD.
Sanitization: There is no evidence of input validation or content filtering for the PRD data before it is processed by the AI.
[DATA_EXFILTRATION]: The 'analyze-complexity' and task 'expand' commands involve remote AI calls, implying that project metadata and requirements are sent to an external service for processing.

cc-taskmaster