The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute multiple shell commands using the akf CLI tool, including akf stamp, akf read, akf inspect, and akf audit. These commands interact directly with the local file system.
[EXTERNAL_DOWNLOADS]: The skill requires the installation of external dependencies: the akf Python package via pip and the akf-format Node.js package via npm. These packages are maintained by an unverified community author and are downloaded from public registries at runtime.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it reads and processes metadata from external files which could contain malicious instructions designed to influence the agent's behavior.
Ingestion points: Metadata is ingested from files via the akf read, akf inspect, and akf audit commands.
Boundary markers: There are no explicit delimiters or instructions provided to the agent to ignore potentially malicious content within the file metadata.
Capability inventory: The skill possesses the ability to read and write to the file system (akf stamp) and execute shell commands.
Sanitization: The instructions do not specify any sanitization, validation, or filtering of the metadata content before it is presented to the agent.

akf-trust-metadata