The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data in the form of issue descriptions, notes, and comments which could potentially contain prompt injection attacks.
Ingestion points: Commands such as bd show, bd list, bd ready, and bd search ingest text from the Beads database.
Boundary markers: The skill does not define specific delimiters or instructions to prevent the agent from following directions embedded within issue content.
Capability inventory: The agent has access to the Bash tool (scoped to bd:*) and file system read access.
Sanitization: No explicit sanitization or filtering of issue data is described in the skill files.
[EXTERNAL_DOWNLOADS]: The documentation provides instructions to download and execute an installation script from the vendor's official GitHub repository.
Evidence: TROUBLESHOOTING.md contains the command curl -fsSL https://raw.githubusercontent.com/gastownhall/beads/main/scripts/install.sh | bash.
Context: The resource originates from the author's verified repository (github.com/gastownhall/beads).
[DATA_EXFILTRATION]: The skill includes functionality to synchronize project data, including audit logs of agent interactions, to remote Dolt servers.
Evidence: The bd sync and bd dolt push commands are used to transmit the local database state to a configured remote.
Context: This is a core feature intended for multi-session and multi-machine collaboration.
[COMMAND_EXECUTION]: The skill requires the Bash tool to interact with the bd CLI.
Constraint: The allowed-tools field in SKILL.md restricts command execution to the bd binary using the bd:* pattern.

beads