The Agent Skills Directory

[SAFE]: The skill provides a comprehensive and structured framework for software estimation using industry-standard techniques like PERT and Fibonacci story points.
[COMMAND_EXECUTION]: The skill utilizes standard Unix utilities (grep, find, wc, git) and the GitHub CLI (gh) to perform read-only analysis of project structure, change history, and PR metadata. These commands are templates intended for local data gathering and do not exhibit signs of persistence, privilege escalation, or malicious intent.
[DATA_EXPOSURE_EXFILTRATION]: Accesses local file contents and repository metadata specifically for complexity analysis and effort calibration. No network operations to unknown or unauthorized domains were detected. The use of the GitHub CLI to list pull requests interacts with a well-known service as expected for the developer use case.
[PROMPT_INJECTION]: The instructions are strictly limited to the estimation process and do not contain directives to bypass safety guidelines, reveal system prompts, or override agent behavior.
[INDIRECT_PROMPT_INJECTION]: The skill defines a surface for processing untrusted data by searching through source code, commit logs, and pull request descriptions.
Ingestion points: Files in src/, git log output, and gh pr list results.
Boundary markers: The instructions define a specific structured output format for the final estimate, providing a clear frame for the agent's response.
Capability inventory: Uses the Bash tool to execute analysis commands.
Sanitization: The skill relies on the agent to treat analyzed content as data to be counted or summarized rather than instructions to be executed.

estimate