The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses a wide array of shell commands (including git, find, grep, sed, awk, and cat) to scan the filesystem, manipulate file contents, and manage repository state. These commands are executed via the setup-team.sh bash script and setup-team.ps1 PowerShell script.
[DYNAMIC_EXECUTION]: Includes inline Python code execution within the bash script (`python3
... <<'PY') to handle JSON processing. It also executes local helper scripts using relative path resolution ($(dirname "$0")/../team-setup/team-helpers.sh`), which is a form of dynamic loading from computed paths.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data during its indexing and validation phases. Specifically:
Ingestion points: Scans .md files in context_modules/, skills/, and .adlc/drafts/ to extract metadata and validate structure.
Boundary markers: The skill does not implement explicit boundary markers or "ignore instructions" delimiters when parsing these external files.
Capability inventory: The skill has broad capabilities including file writes, shell command execution, and Git operations (git push).
Sanitization: There is no evidence of input sanitization or escaping of the content read from these markdown files before they are processed or used to generate new artifacts.
[DATA_EXFILTRATION]: The setup-team.sh script performs a git push origin "$adlc_branch" operation. While this is a functional requirement for synchronizing reports and confidence scores with the remote repository, it constitutes a network operation that transmits local data to an external server.

team-repair