The Agent Skills Directory

[SAFE]: The skill configuration employs security best practices by restricting the Bash tool to a specific, benign command (date) used for generating timestamps in metadata. This minimizes the attack surface for command injection.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface because it processes untrusted user input (transcripts) to generate structured PRDs and create project management tasks.
Ingestion points: Data provided by the user via the $ARGUMENTS variable, which can be raw text or a file path (processed in SKILL.md).
Boundary markers: Absent; the skill does not use specific delimiters to isolate the content of the transcript from the agent's core instructions.
Capability inventory: The skill can perform file system writes (Write) to the docs/prd/ directory and create/update projects and issues via the Linear MCP server integration.
Sanitization: No explicit sanitization or filtering logic is present to identify or ignore instructions that may be embedded within the user-provided transcript.
[EXTERNAL_DOWNLOADS]: The README.md file contains a reference to an external repository on GitHub (github.com/ibraheem4/linear-mcp) for users who wish to install the optional Linear MCP server. This is a documentation link to a well-known service and does not result in any automated downloads or code execution during the skill's operation.

create-prd