ctx-check

A codebase-checking command that compares claims in AI instruction and memory files with the files and structure that actually exist in a project.

In plain words
What is it for?
Use it to check files such as CLAUDE.md or GEMINI.md, validate selected files against each other, or review context files for a particular AI tool.
Why use it?
AI context files can become outdated or disagree with the code, causing coding agents to follow incorrect instructions. This command identifies stale, broken, or conflicting references.

Command

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 commands/knitli/toolshed/ctx-check
Clone the repo
git clone --depth 1 https://github.com/knitli/toolshed
Per session 21 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 653 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 $0.00021 $0.00653
Opus 5 $0.00010 $0.00327
Sonnet 5 $0.00004 $0.00131
Haiku 4.5 $0.00002 $0.00065

Measured 2d ago against content hash d12e805696e2, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

ctx-check 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/ctx/commands/ctx-check.md · 48 lines

How it starts

The opening of the file, as written. The whole thing — 48 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/ctx:check — Staleness Validation

Validate claims in AI context files against the actual codebase. Find what's stale, what's broken, and what's drifted from reality.

Arguments

$ARGUMENTS

  • No arguments: discover all context files, then validate every memory/instruction file.
  • File paths (e.g., /ctx:check CLAUDE.md): validate only those specific files.
  • "validate X against Y" or "X against Y" (e.g., /ctx:check CLAUDE.md against GEMINI.md): extract claims from all named files, then cross-validate — check whether the claims in each file hold true not just against the codebase but also against the other named files. Flag any disagreements as drift alongside staleness.
  • Ecosystem names (e.g., /ctx:check cursor): validate all context files belonging to those ecosystems.

Instructions

First, run discovery (same as /ctx:discover) to identify all context files — unless the user named specific files above, in which case use those directly. Then, for each memory and instruction file in scope (skip config-only files and old planning output), do the following:

Extract and validate claims

Read each file and extract every factual claim that can be checked against the codebase:

  1. Paths: Any file or directory path in backticks or quotes. → Does it exist?
  2. Versions: Semver strings, edition years, rust-version values. → Compare against manifest files (Cargo.toml, package.json, pyproject.toml, go.mod).
  3. Dependencies: Named packages with version constraints. → Compare against manifests and lock files.
  4. Symbols: Type names, function names, module names, crate names in backticks. → Search the codebase for their existence.
  5. Counts: "N crates", "supports N languages", "seven modules". → Count the actual items.
  6. Commands: Build/test/lint commands in code blocks. → Check task runner configs (mise.toml, Makefile, package.json scripts, justfile).
  7. Technology claims: "uses X", "built with Y", "powered by Z". → Verify the dependency/tool is actually present.

Read the full file on GitHub · 48 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 · 48 lines · 21 tokens per session scan A d12e805696e2

Subscribe to this mod's changes

ctx-check is a command published in the GitHub repository knitli/toolshed (1 stars, last pushed 4mo ago), licensed MIT. It adds 21 tokens to every session and 653 once invoked, about $0.0001 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.