The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to aggregate content from external source files into a single context bundle to be used as a prompt. It lacks sanitization or boundary enforcement, which could lead to an agent following malicious instructions embedded within the source files.
Ingestion points: The scripts/context_bundler.py and scripts/token_counter.py scripts read the contents of files specified via command-line arguments or configuration files.
Boundary markers: The tool uses markdown headers (e.g., ## DATA SCHEMA, ## BUSINESS CONTEXT) to separate sections but does not provide explicit 'ignore instructions' markers for the ingested content.
Capability inventory: The skill possesses file read and write capabilities on the local file system.
Sanitization: No filtering, escaping, or validation is performed on the content of the read files before they are concatenated into the final bundle.
[COMMAND_EXECUTION]: The skill relies on the execution of Python scripts (scripts/context_bundler.py and scripts/token_counter.py) that perform file system operations based on user-provided input paths.

context-packager