The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local CLI commands to interact with the environment, including openspec, git, and mkdir. These commands are standard for the skill's purpose of managing local development workflows.
[DATA_EXPOSURE_AND_EXFILTRATION]: The skill performs local analysis of the codebase by scanning for TODO comments, console.log statements, and testing coverage. This data is used within the session to suggest improvements and no evidence of external data transmission (exfiltration) was found.
[INDIRECT_PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it ingests untrusted content from the codebase (e.g., source code comments and file paths) during the initial scanning phase.
Ingestion points: Reads local codebase content (comments, console logs, file names).
Boundary markers: No explicit delimiters or warnings to ignore instructions found within the scanned code are present.
Capability inventory: Executes shell commands (openspec, git, mkdir) and writes files to the local file system.
Sanitization: There is no explicit logic described to sanitize or escape data extracted from the codebase before it is used to generate task descriptions or command arguments.
[DYNAMIC_EXECUTION]: The skill generates kebab-case names for new development 'changes' based on user-selected tasks. While this involves dynamic string assembly, it is used for local file naming and CLI arguments within the OpenSpec framework.

openspec-onboard