The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill facilitates the installation of the fallow CLI tool from official and well-known registries including NPM and Crates.io. These are standard procedures for developer tooling and utilize trusted services.
[COMMAND_EXECUTION]: The skill involves the execution of various shell commands to perform static analysis. It includes critical safety instructions for the agent, such as using the --format json flag for machine-readable output and appending || true to prevent command chain failures. It also provides features like git hook initialization which are standard in development workflows.
[PROMPT_INJECTION]: This skill possesses an indirect prompt injection surface as it processes untrusted codebase data. However, it effectively mitigates this risk through several built-in strategies.
Ingestion points: Local source code files (.ts, .js), configuration files (JSON/TOML), and git history via the shell commands.
Boundary markers: The agent is explicitly instructed to use the --format json flag and redirect stderr to /dev/null, ensuring that progress messages and potential malicious content in the code do not interfere with the agent's parsing logic.
Capability inventory: The skill allows file system modifications via the fallow fix command and executes various analysis subcommands. The agent is instructed to perform dry-runs and seek user confirmation before applying fixes.
Sanitization: The skill relies on the static parsing capabilities of the fallow tool and follows the principle of least privilege by scoping output via filter flags.

fallow