Results for “tool-abuse”

Conduct authorized security assessments of LLM applications and AI agents, covering prompt injection, tool abuse, RAG exposure, memory poisoning, and model supply-chain risks.

Apply least-privilege tool allowlisting, identity binding, and human-in-the-loop controls for agent tool calls.

Design and validate model-facing tool definitions with clear names, action-oriented descriptions, bounded JSON Schema parameters, explicit side effects, safe defaults, idempotency, errors, and realistic tests. Use when creating function-calling tools, MCP tools, agent actions, structured tool inputs, or when a model selects the wrong tool, invents arguments, or causes unsafe side effects.

Design tool descriptions, schemas, and response formats that agents can reliably route on, including MCP server design, tool-set consolidation, and naming conventions.

Scan Model Context Protocol servers and tool metadata for poisoning, SSRF, and unauthenticated exposure.

Audits ECC Tools GitHub App for cost issues like runaway PR creation, quota bypass, premium-model leakage, and duplicate jobs, using an evidence-first workflow.

Pre-install security scanner for AI agent skills. 7.5% of 14,706 skills are malicious. Audit before you trust.

Agent skill for user-tools - invoke with $agent-user-tools

Creates or modifies agent tools for OpenAI Agents SDK, covering search, analysis, API integration, and database query patterns with caching and metrics.

Compacts the current conversation into a handoff document for another agent to pick up, referencing existing artifacts by path or URL instead of duplicating them.

Fetches and analyzes LangSmith execution traces to debug LangChain and LangGraph agents, investigating errors, tool calls, and performance.

Use with analysis-agent or task-agent for task-local transaction, isolation, and conflict decisions. Do not use without a transaction decision or as task owner.

Audit agent skills, plugins, prompts, manifests, scripts, dependencies, and bundled assets for provenance, prompt-injection, permission, execution, exfiltration, persistence, and update risk. Use when evaluating a third-party skill before installing, enabling, updating, publishing, or distributing it; reviewing an untrusted SKILL.md, agent configuration, MCP integration, archive, or repository; comparing a package with a known-good version; or investigating unexpected tool, network, credential, or filesystem behavior.

Compact the current conversation into a handoff document for another agent to pick up, including suggested skills and references to existing artifacts.

Sets up and audits the Claude Code harness for a project: CLAUDE.md, .claude/rules, skills, subagents, settings.json permissions, hooks, verification loop. Commands: kickoff (scaffold new setup), check-playbook (score an existing one), scouting-report (last scorecard), adjust (fix bloat/misplaced instructions), drill (turn advisory prose into real hooks/permissions/CI), decline (record an accepted risk), gameplan (work order with acceptance criteria before building), watch-film (check a diff against that order for scope creep/weakened tests/false claims). Use for setting up Claude Code, or on: "Claude ignores my CLAUDE.md", "it's huge and still misses things", "it said done but ran nothing", "it changed files I didn't ask about", "it weakened a test to pass", "rule, skill, or hook?", "is my setup any good". Not for code quality (code-audit), test coverage (test-assessment), one-off prompt wording (genie-proof-prompts), new skill authoring (skill-creator), or compacting a conversation (handoff).

`analysis-agent`/`task-agent`/`review-agent`: use for changed assets, trust boundaries, reachable abuse paths, impact, or control placement; skip without a security delta.

Use with task-agent or review-agent for task-local cache scope, freshness, invalidation, and source-load risk. Do not use without a cache decision or as task owner.

Report AI token usage and estimated costs. Use when: owner asks about costs today/yesterday/this week, per session, or per model. Shows main session, cron jobs, and subagents. Answers: 'how much did today cost?', 'how much was this session?', 'what was last week's spend?'

Inspect and adjust the agent's runtime state, including model, context window, iteration limits, token usage, workspace configuration, subagent status, and request routing metadata.

Use `analysis-agent` to analyze permissions, secrets, sensitive data, trust boundaries, and injection; `task-agent` to implement controls; and `review-agent` to assess evidence. Skip self-review and no-trust-impact work.

`analysis-agent`/`task-agent`/`review-agent`: use when Rust changes cross ownership, unsafe/FFI, panic, cancellation, or Send/Sync boundaries; skip tool-only work.

Agent skill for issue-tracker - invoke with $agent-issue-tracker

`analysis-agent`/`task-agent`: use when actor journeys change entries, branches, interruption, recovery, authority, or side effects; skip route/state-only work.

Use with analysis-agent, task-agent, or review-agent for task-local authentication lifecycle and recovery risk. Do not use without that decision or as task owner.

Analyzes prompt injection, retrieval poisoning, memory contamination, planner drift, and tool-boundary abuse in agentic systems, mapping trust boundaries and proving exploit chains.

Detect model stealing, model inversion, and membership inference performed through inference-API abuse by monitoring query patterns, applying output perturbation, and red-teaming your own model's extractability.