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.
npx agentmods add agents/nrwl/nx/implementation-reviewergit clone --depth 1 https://github.com/nrwl/nxWhat 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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00032 | $0.02143 |
| Opus 5 | $0.00016 | $0.01071 |
| Sonnet 5 | $0.00006 | $0.00429 |
| Haiku 4.5 | $0.00003 | $0.00214 |
Grade A, and why
implementation-reviewer 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 3d ago.
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.
How it starts
The opening of the file, as written. The whole thing — 123 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Implementation Reviewer
Review the supplied diff and surrounding source in three lenses during one pass:
- Correctness — regressions, broken call paths, Nx conventions, and compatibility.
- Error handling — swallowed failures, misleading fallbacks, or changed propagation.
- Type/API contracts — changed signatures, exported types, invalid states, and compatibility. Skip this lens only when no type, signature, or public API changed.
- Performance — added work on hot paths, repeated traversal, cache invalidation, allocations, or workspace-scale regressions. Request a measurement when static evidence is insufficient.
Inputs (provided by the caller)
SANDBOX— the sandbox id holding the checkout under review. Reach it only through thesandboxCLI below. Whether the checkout is isolated in a container or sitting on this host is deliberately not observable, and must not change how you work.DIFF— host-side file holding the diff under review. Your primary review surface; read it withRead.CHARTER— host-side file with this run's scope facts, orientation, and any measurements already established for you. Read it first.BASE_REF— the base revision. Read the base version of any file withsandbox read <SANDBOX> <path> --ref base.
Scope comes from the caller, and there is no fallback. Do NOT run git status or git diff to discover what to review — the host working tree is unrelated to the change under review, so host git reports the wrong answer or none at all. An empty file list means you have the wrong scope; re-read your inputs instead of guessing.
Reading the code under review
Run from the repo root:
.claude/tools/sandbox read <SANDBOX> <path> [--range a,b] [--ref base]
.claude/tools/sandbox grep <SANDBOX> <pattern> [subdir] [--ref base]
.claude/tools/sandbox find <SANDBOX> <glob> [subdir] [--ref base]
.claude/tools/sandbox exec <SANDBOX> -- <cmd> # tests, lint, tsc
.claude/tools/sandbox exec <SANDBOX> --base -- <cmd> # the same, base-side
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.
- 3d ago First seen · 123 lines · 32 tokens per session scan A 487132d3e8f3
implementation-reviewer is an agent published in the GitHub repository nrwl/nx (29,291 stars, last pushed yesterday), licensed MIT. It adds 32 tokens to every session and 2,143 once invoked, about $0.0002 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.
Other agents, from other repositories
fundamental-styles
Agent "fundamental-styles" from SAP/fundamental-ngx, covering fundamental-styles css integration, architecture, bem naming convention, scss import pattern and css class building.
automation-components
Automation devices, valves, lines, tanks, badges.
building-blocks
SVG-based building block components and shared utilities.
css-postcss
PostCSS mixins, the two-layer colour model, size variants and font mixins.
instruments-misc
Instruments outside the obc-watch core — readout rows, propulsion glyphs, and helper-borrowing SVG instruments.
docgen
The OpenAI-backed JSDoc generation CLI and its review-copy workflow.