The Agent Skills Directory

[SAFE]: The skill provides extensive instructional content for Y Combinator (YC) resources and application processes. It references several well-known and official domains such as ycombinator.com, startupschool.org, workatastartup.com, and news.ycombinator.com. These references are entirely appropriate for the skill's function.
[COMMAND_EXECUTION]: The skill includes a Python script (scripts/check-output.py) intended to validate deliverables by checking for the presence of specific keywords. The script's operations are benign, limited to reading the text of a provided file and printing a pass/fail status based on keyword counts.
[INDIRECT_PROMPT_INJECTION]: The skill demonstrates an attack surface for indirect prompt injection as it ingests user-supplied content for validation. However, the associated risk is minimal because the provided validation script lacks dangerous capabilities such as network access, file writing, or code execution based on the input.
Ingestion points: The validation script scripts/check-output.py reads the contents of a user-specified file.
Boundary markers: The SKILL.md file contains a quality check section that provides instructions for verification.
Capability inventory: The skill utilizes a basic keyword matching script with no network, write, or dynamic execution permissions.
Sanitization: No sanitization is implemented for the text read by the validation script.

yc-ecosystem