The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes shell commands via the Bash tool to perform Git operations such as checkout, pull, branch, add, commit, push, log, and diff, as well as GitHub CLI operations like gh pr create and gh pr view. These commands are used to manage the development lifecycle.\n- [DATA_EXFILTRATION]: The skill performs network operations by pushing code to a remote repository (git push) and creating pull requests via the GitHub CLI. These operations target github.com, which is a recognized domain for development workflows.\n- [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its processing of external data.\n
Ingestion points: The skill accepts user arguments ($ARGUMENTS), captures user feedback, and reads data from files like ROADMAP.md and sprint plan artifacts (Step 1.2, Step 3.1, Step 4.2 in SKILL.md).\n
Boundary markers: No explicit boundary markers or delimiters are used to isolate untrusted data from instructions.\n
Capability inventory: The skill has access to shell execution (Git, GH CLI), file writing, and network operations across all phases.\n
Sanitization: There is no evidence of sanitization or validation of external content before it is interpolated into shell commands or report templates.\n- [CREDENTIALS_UNSAFE]: The skill incorporates a security baseline check (Step 0.4 in SKILL.md) that scans for patterns indicative of hardcoded secrets (e.g., sk-, ghp_, token=) using grep before any code is committed. This serves as a defensive measure to prevent accidental credential exposure.

team-sprint