The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/list_azdo_repos.py utilizes subprocess.run to interact with external tools required for its functionality.
It executes az (Azure CLI) commands to list projects, fetch descriptions, and list repositories within an Azure DevOps organization.
It executes git remote get-url origin to identify the origin of local directories and map them to remote repositories.
[INDIRECT_PROMPT_INJECTION]: The skill processes repository names and project descriptions fetched from Azure DevOps, which are then written to the ALL_REPOS.md file for agent consumption.
Ingestion points: Data is ingested in scripts/list_azdo_repos.py via list_projects, project_description, and list_repos functions.
Boundary markers: The output is structured as a Markdown table in ALL_REPOS.md, but lacks specific instructions for the agent to ignore potentially malicious content embedded in the descriptions.
Capability inventory: The skill performs filesystem reads via os.walk, file writes to ALL_REPOS.md, and system command execution via subprocess.run.
Sanitization: The script does not sanitize or filter the content of project/repository descriptions for natural language instructions before writing them to the summary file.

cross-repo-discovery