The Agent Skills Directory

[COMMAND_EXECUTION]: The skill makes extensive use of the paperclipai command-line interface to read and update organization state. It also utilizes shell commands like curl and command -v for environment verification and API interactions.
[EXTERNAL_DOWNLOADS]: The instructions direct the user to install the paperclipai CLI via npm install -g and execute the @bbengamin/paperclip-mcp-server using npx. These are identified as vendor resources provided by the skill author.
[PROMPT_INJECTION]: The skill is subject to indirect prompt injection as it ingests untrusted data from Paperclip issue descriptions, comments, and wiki pages.
Ingestion points: Reads issue metadata, descriptions, plan documents, and external wiki content via paperclip-wiki-fetch (documented in SKILL.md and cli-contract.md).
Boundary markers: No explicit content delimiters are defined for the ingested data.
Capability inventory: The skill can mutate issue status, assign agents, update documents, and perform REST API requests (documented in integration-matrix.md).
Sanitization: No explicit sanitization of ingested content is mentioned.
Mitigation: The risk is significantly mitigated by a mandatory 'Operator Approval Boundary' and 'Mutation Rule' that require the agent to present all proposed changes for human approval before execution.

paperclip-triage