The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The instructions in SKILL.md direct the agent to use npx @superdoc-dev/cli@latest. This command fetches and executes code from the public NPM registry at runtime. The use of the @latest tag ensures that the agent always runs the most recent version of the package, which is a significant security risk as the package content can be changed by the maintainer at any time without user review, potentially leading to the execution of malicious updates.
[EXTERNAL_DOWNLOADS]: The skill relies on fetching the @superdoc-dev/cli package from an external source (the NPM registry) during execution.
[COMMAND_EXECUTION]: The skill facilitates the execution of shell commands through the npx utility to perform operations on local Word documents.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it reads and processes data from external Word documents which could contain malicious instructions.
Ingestion points: Text content from .docx files is ingested via the read and search commands defined in SKILL.md.
Boundary markers: Absent. There are no instructions or delimiters provided to help the agent distinguish between its own instructions and the content retrieved from documents.
Capability inventory: The skill is capable of shell execution via npx.
Sanitization: Absent. The skill does not define any validation or filtering mechanisms for the data extracted from the documents.

docx-processing-superdoc