Statically review a Claude Code / AI-agent setup for security misconfigurations — risky hooks, over-broad permissions, untrusted or cleartext MCP servers, hardcoded secrets, endpoint redirection, over-privileged agents/skills, and prompt-injection in steering files. Use when asked to audit a `.claude/` directory, `settings.json`, `.mcp.json`, a plugin/marketplace, or `CLAUDE.md`, or to add a config-security gate to CI. Drives the `agentscanner` CLI.
Review the training / MLOps pipeline for security weaknesses — data-poisoning surface, feature-store and data-source trust, experiment-tracking and model-registry access control, secrets, and reproducibility/provenance. Use when assessing how models are built and promoted, not how they're served or consumed.
Review the provenance and integrity of models, datasets, and ML artifacts pulled from hubs/registries — unsafe deserialization (pickle, PyTorch/Keras/joblib), untrusted model sources, model/dataset tampering, and signing. Use when an app loads third-party model weights or datasets, or before promoting a model.
Harden a deployed model inference endpoint — authn/authz, rate limiting and abuse control, input validation, and exposure to model-extraction / inversion / membership-inference attacks. Use when reviewing how a model is served (REST/gRPC, Triton/TorchServe/KServe, or a hosted inference API), not how it was trained.
Export a finished security artifact (report, executive summary, diagram, spreadsheet of findings) to Google Drive — right folder, clear naming, and shareable with the intended audience. Use when a deliverable needs to be handed to stakeholders who live in Drive/Docs/Sheets rather than a wiki or tracker.
Turn a security finding (vuln, pentest issue, review item) into a well-formed Jira issue — mapped severity→priority, remediation/repro in the body, labels and components set, and dedup-checked against existing issues. Use when findings need to become tracked, assignable work in Jira.
Publish a security report, runbook, threat model, or assessment writeup to Confluence — correct space/parent, consistent page structure, labels, and links back to related issues/pages. Use when a finished document needs to live in the team wiki, not just a local file.
Look up MITRE ATT&CK tactics, techniques, and mitigations, and map an observed behavior, finding, or detection to the right technique ID(s). Use whenever work needs a consistent ATT&CK reference — detection coverage, threat reports, red-team TTP planning, or tagging a finding.