--- name: agent-feature-review description: >- Guided security review of a new agent feature being switched on for a model or product the organisation already uses, when someone asks "can we enable this?" and no single benchmark answers it. One threat-and-test template per feature type: computer use and browser control, persistent memory, connectors and MCP servers, code execution sandboxes, file search and RAG, multi-agent and subagents, scheduled and background agents. Each template says what the feature changes about the trust boundary and, per threat, the review questions, the vendor evidence to look for, which existing benchmark could test it and through which skill, and what has no standardized test. Threats carry OWASP LLM 2025, OWASP Agentic and MITRE ATLAS ids where verified. Writes a JSON and markdown record, evidence.jsonl rows, and a list of every threat left unassessed. It prescribes no control and makes no recommendation. For an image-generation model use image-model-review. license: MIT compatibility: "Python 3.10+ standard library only. No API key and no network access are needed by the scripts." --- # agent-feature-review — security review of a new agent feature You are running this skill **on behalf of a reviewer** whose organisation is about to turn on an agent feature. You pick the template for the feature, capture the deployment context, walk the reviewer through each threat's questions, record the evidence they bring, and write a record that separates what was examined from what was not. Three rules hold throughout: - **You record. You do not decide.** No rating, no "approve", no bar. The decision block is written empty. - **You do not prescribe controls.** Record the controls that are in place and which were examined. Do not propose ones to add. - **Nothing is dropped silently.** Every threat in the template ends up in the record as `evidence_recorded`, `answers_only`, `not_applicable` (with a reason) or `not_assessed`. ## What the reviewer supplies | Input | Used for | |---|---| | The feature type, and the exact product, vendor, model and feature version | choosing the template; the record's subject | | The deployment context: what the agent can touch, who starts a run, where a person approves, whose identity it uses | the `deployment` block (`references/deployment-context.md`) | | Vendor documents for this feature: system card, product and security documentation, data-handling terms | `vendor_evidence` and `vendor_absent` per threat | | Results of any tests already run, with the path of their summary file | `tests` per threat | | What they checked themselves in a console, a configuration or a contract | `reviewed` per threat | What they cannot supply stays empty and is reported as not assessed. ## Steps Work in a review folder. If this is part of a wider review, use that review's folder and pass its `evidence.jsonl` with `--evidence`; otherwise create `.evals/agent-feature-review/