The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user or agent to install a CLI tool using npx -y @mvanhorn/printing-press and go install github.com/mvanhorn/printing-press-library/.... These resources are hosted on NPM and GitHub under the skill author's namespace.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests unstructured data from external YesWeHack API endpoints, such as public hacktivity writeups and program rules. It specifically provides recipes that pipe this untrusted data into an LLM for summarization (e.g., hacktivity learn ... | claude).
Ingestion points: Data enters via the yeswehack-pp-cli commands that fetch program details (programs get) and disclosed reports (hacktivity learn).
Boundary markers: The tool uses a --agent flag to produce structured JSON, but the skill instructions do not provide delimiters or warnings to the LLM regarding instructions potentially embedded within the bug reports themselves.
Capability inventory: The agent can execute shell commands via the Bash tool, write report drafts to the local filesystem, and submit reports to the platform.
Sanitization: There is no documented sanitization or filtering of the external report content before it is passed to the LLM context.
[COMMAND_EXECUTION]: The skill relies on executing the yeswehack-pp-cli binary using the Bash tool for all its core functions.
[DATA_EXFILTRATION]: The CLI tool includes a --deliver webhook:<url> feature that allows it to POST command output directly to a remote URL. While presented as a data-routing feature for agents, this mechanism could be misused to exfiltrate sensitive bug bounty information, such as private program scopes or report drafts, to an unauthorized endpoint.

pp-yeswehack