The Agent Skills Directory

[SAFE]: Extensive review of the skill's instructions, examples, and reference files reveals no malicious behavior. The skill provides clear, benign guidelines for professional outreach.
[COMMAND_EXECUTION]: The skill instructions in references/banned-patterns.md mention the execution of a local script, writer/scripts/lint.py, using the Bash tool to validate email drafts. This script is an internal component of the vendor's toolset and does not involve untrusted remote code execution.
[PROMPT_INJECTION]: The skill is subject to indirect prompt injection as it processes user-supplied data such as recipient context and previous email threads. Ingestion points: Inputs for recipient context, ask, proof, and previous email content provided as arguments to the modes defined in SKILL.md. Boundary markers: No explicit delimiters or boundary markers are defined in the instructions to isolate untrusted user data from the system's instructions during prompt assembly. Capability inventory: The agent has access to Bash, Write, Edit, Read, Grep, and Glob tools as specified in the SKILL.md frontmatter. Sanitization: The skill relies on tone and structural linting via the lint.py script rather than security-focused sanitization or filtering of external content.

cold-email