auditing-agentic-systems

A checklist-based reviewer for AI agents, their tool-use loops, and requests sent to language models.

In plain words
What is it for?
Use it to inspect an agent definition, code, prompt, tool loop, or written system description and produce scored findings with fixes.
Why use it?
It reveals gaps in prompts, tools, context, caching, evaluation, and safety controls before they cause problems.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/agentic-dev3o/devx-plugins/audit
Any agent
npx skills add agentic-dev3o/devx-plugins --skill audit
Clone the repo
git clone --depth 1 https://github.com/agentic-dev3o/devx-plugins

Made for: Claude Code, Codex.

Per session 118 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,470 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00118 $0.01470
Opus 5 $0.00059 $0.00735
Sonnet 5 $0.00024 $0.00294
Haiku 4.5 $0.00012 $0.00147

Measured yesterday against content hash fd19610d0a43, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

auditing-agentic-systems scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

plugins/agentic-engineering/skills/audit/SKILL.md · 110 lines

How it starts

The opening of the file, as written. The whole thing — 110 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Agentic System Audit

Target: $ARGUMENTS (path to file, directory, prompt text, or description of the system to audit)

Workflow

Progress checklist:

Agentic System Audit:
- [ ] Step 1: Resolve audit target
- [ ] Step 2: Detect framework and identify components
- [ ] Step 3: Evaluate across six dimensions
- [ ] Step 4: Classify findings and assign severity
- [ ] Step 5: Produce scored gap analysis

Step 1: Resolve Audit Target

The target is one of:

  • Agent definition file — markdown with frontmatter, JSON/YAML config, or code (TS/Py)
  • Inference call site — code that builds messages and calls an LLM API
  • Tool loop — code that iterates on tool calls until done
  • Prompt text — a system prompt, user prompt template, or both
  • Description — the user describes the system in natural language

If $ARGUMENTS is a path, read the file(s). If a directory, list it and ask the user to scope to specific files when more than ~5 are present. If it is prompt text, treat the text as the target directly. If it is a description, ask one clarifying question only when the framework or component shape is genuinely ambiguous.

Step 2: Detect Framework and Identify Components

Identify the runtime context to ground recommendations:

  • Anthropic SDK (anthropic, @anthropic-ai/sdk) — supports prompt caching, adaptive thinking, effort parameter, parallel tool use
  • OpenAI SDK (openai) — tool calling, structured outputs, parallel calls, no native prompt caching as of writing
  • Vercel AI SDK (ai, @ai-sdk/*) — provider-agnostic tool calling, streaming, structured generation
  • OpenAI Agents SDK (@openai/agents) — handoffs, guardrails, tracing
  • Custom / framework-less — raw HTTP calls, hand-rolled loops

Identify the components present in the target: system prompt, user prompt, tool definitions, tool loop, memory/state, evals, guardrails, retries, streaming.

Step 3: Evaluate Across Six Dimensions

For each dimension, load the matching reference and walk the checklist against the target. Do not declare a finding without concrete evidence (a quote from the prompt, a line in the code, or a missing pattern that should be present).

Read the full file on GitHub · 110 lines

Files

What ships with it

7 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday First seen · 110 lines · 118 tokens per session scan A fd19610d0a43

Subscribe to this mod's changes

auditing-agentic-systems is a skill published in the GitHub repository agentic-dev3o/devx-plugins (11 stars, last pushed 14d ago), licensed MIT. It adds 118 tokens to every session and 1,470 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other skills, from other repositories

systematic-debugging

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.

obra/superpowers · 21 tokens

next-cache-components-adoption

Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…

vercel/next.js · 95 tokens

babysit-pr

Babysit a GitHub pull request after creation by continuously polling review comments, CI checks/workflow runs, and mergeability state until the PR is merged/closed or user help is required. Diagnose failures, retry likely flaky failures up to 3 times, auto-fix/push branch-related issues when appropriate, and keep…

openai/codex · 114 tokens

imagegen

Generate or edit raster images when the task benefits from AI-created bitmap visuals such as photos, illustrations, textures, sprites, mockups, or transparent-background cutouts. Use when Codex should create a brand-new image, transform an existing image, or derive visual variants from references, and the output…

openai/codex · 113 tokens

cpu-profile-analysis

Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…

microsoft/vscode · 71 tokens

next-cache-components-optimizer

Drive a Next.js route to instant navigation by setting up an agentic loop, under Cache Components / PPR, on initial load (hard navigation) and client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route at a time; the shipped test then…

vercel/next.js · 170 tokens