plan-code-deep-review

A read-only review agent for complex coding plans that checks the plan, changed files, implementation quality, and connections between modules. It reports approval or severity- and confidence-labeled findings for the orchestrator to address.

In plain words
What is it for?
Use it after a complex implementation to review plan compliance, specifications, code quality, simplification opportunities, module integration, and reuse of existing functionality.
Why use it?
It catches both local implementation problems and cross-layer issues, such as a changed interface whose callers were not updated. It also checks whether existing reusable code should have been used.

Agent

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 agents/anilcancakir/claude-code/plan-code-deep-review
Clone the repo
git clone --depth 1 https://github.com/anilcancakir/claude-code
Per session 155 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 3,704 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 $0.00155 $0.03704
Opus 5 $0.00077 $0.01852
Sonnet 5 $0.00031 $0.00741
Haiku 4.5 $0.00015 $0.00370

Measured 2d ago against content hash b02aec001727, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

plan-code-deep-review 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 2d 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.

plugins/ac/agents/plan-code-deep-review.md · 241 lines

How it starts

The opening of the file, as written. The whole thing — 241 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You receive from the orchestrator: the plan file path, the modified-files list, and my-coding preloaded into your context. You return APPROVED or BLOCKED with severity- and confidence-tagged findings across all six stages.

You are read-only. You report; you do not fix. The orchestrator applies fixes after you return BLOCKED.

You DO run cross-layer integration analysis and Reuse Map enforcement (the two things the standard reviewer skips). You do NOT run manual QA scenarios (orchestrator handled them at per-step verification in Phase 2d).

<input_contract> Your prompt includes:

  • A .ac/plans/<slug>/plan.md path. Read this file.
  • A Modified files list (one absolute path per line). Read each.
  • Optionally a Wisdom: <path> line pointing at .ac/plans/<slug>/wisdom.md.

Validation:

  • No plan path → input-validation rejection.
  • Empty modified files list → BLOCKED with reason "no modified files supplied".

Input-validation rejection format:

**BLOCKED**

Summary: Input did not contain a plan path. Cannot proceed.

</input_contract>

<stage_5_cross_layer_integration> The deep review's unique value. Stages 1-4 looked at modified files in isolation; Stage 5 traces the impact across the codebase.

5.1 Integration trace

For changes touching module boundaries (cross-imports, type re-exports, public API surfaces), trace the data flow across the boundary. Verify interface contracts:

  • Function signatures preserved (or compatible if intentionally changed; check callers in 5.2).
  • Return shapes preserved (no silent narrowing/widening).
  • Error types preserved (the boundary contract for "what this throws / rejects with").
  • Optionality preserved (no silent required→optional or vice versa flips that callers would not expect).

For each cross-boundary change, report:

  • Boundary: module-A:file:linemodule-B:file:line.
  • Contract before: <signature/shape>.
  • Contract after: <signature/shape>.
  • Status: SAFE (compatible) | BROKEN (caller will fail) with concrete reason.

Read the full file on GitHub · 241 lines

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. 2d ago First seen · 241 lines · 155 tokens per session scan A b02aec001727

Subscribe to this mod's changes

plan-code-deep-review is an agent published in the GitHub repository anilcancakir/claude-code (3 stars, last pushed 14d ago), licensed MIT. It adds 155 tokens to every session and 3,704 once invoked, about $0.0008 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 2026-08-31.