The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process text corpora provided by users, creating an attack surface for indirect prompt injection where malicious instructions embedded in the data could influence the agent's actions.
Ingestion points: User-provided text corpora (referenced as docs or texts in SKILL.md and references/text_stack.md).
Boundary markers: Absent; there are no instructions to use delimiters or warnings to ignore embedded instructions in the processed data.
Capability inventory: The skill has the ability to execute Python code via the Bash tool to perform analysis.
Sanitization: Absent; no sanitization or validation of the input text is present in the provided scripts or templates.
[COMMAND_EXECUTION]: The skill instructs the agent to execute generated Python code snippets to perform text analysis tasks such as topic modeling and clustering.
Evidence: SKILL.md configures allowed-tools: Read Bash(python:*) and references/text_stack.md provides multiple code blocks for execution via a Python interpreter.
[EXTERNAL_DOWNLOADS]: The skill utilizes the sentence-transformers library, which fetches pre-trained model weights from Hugging Face's official model registry.
Evidence: references/text_stack.md contains code calls like SentenceTransformer("all-MiniLM-L6-v2") which triggers weight downloads.

alterlab-text-as-data