02-context-grounding

02-context-grounding is a cursor rule for coding agents from AG-Dmitry/composer-enhanced. It costs 24 tokens per session (577 once invoked), scanned A, original, MIT.

A preparation step that gathers verified information about the code before any edits. It checks that files, APIs, names, imports, and paths actually exist.

In plain words
What is it for?
Use it at the start of a change to read the files involved, confirm external references, and trace how related code is used.
Why use it?
It reduces mistakes caused by guessing how a library or project is structured.

Cursor rule

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 rules/ag-dmitry/composer-enhanced/02-context-grounding
Clone the repo
git clone --depth 1 https://github.com/AG-Dmitry/composer-enhanced

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for 02-context-grounding

README.md
[![agentmods](https://agentmods.dev/badge/rules/ag-dmitry/composer-enhanced/02-context-grounding.svg)](https://agentmods.dev/rules/ag-dmitry/composer-enhanced/02-context-grounding)
Your own site
<a href="https://agentmods.dev/rules/ag-dmitry/composer-enhanced/02-context-grounding"><img src="https://agentmods.dev/badge/rules/ag-dmitry/composer-enhanced/02-context-grounding.svg" alt="Measured on agentmods" height="20"></a>
Per session 24 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 577 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 $0.00024 $0.00577
Opus 5 $0.00012 $0.00289
Sonnet 5 $0.00005 $0.00115
Haiku 4.5 $0.00002 $0.00058

Measured 4d ago against content hash 6c7b8cd84812, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

02-context-grounding 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 4d 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.

agent-protocols/02-context-grounding.mdc · 60 lines

How it starts

The opening of the file, as written. The whole thing — 60 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Phase 2 — Context Grounding

Persona: Planner. No code in this phase.

Cheap models hallucinate APIs, import paths, and file contents. This phase replaces guessing with evidence.

The anti-assumption rule

If you have not read it, you do not know it. Anything unverified is an assumption, and assumptions must be checked or asked — never coded against.

Steps

  1. Read the files you will touch, fully enough to understand them. Never edit a file you have not read in this session.

  2. Verify every external reference you intend to use actually exists:

    • Functions / hooks / components — confirm the real signature, not a guessed one.
    • Import paths and module exports — confirm they resolve.
    • Types and fields — confirm they exist with the names/shapes you expect. Use search/read tools. Do not rely on memory of "how this library usually works".
  3. Trace the data flow / call sites relevant to the change, so you know what you might break.

  4. Keep context high-signal (avoid context rot). Read widely to find the relevant slice, but keep only high-signal context in view — a cheap model's attention degrades and cost rises when the window fills with irrelevant tokens. For broad exploration, prefer a readonly explore subagent so only its digest enters your context (see escalation-subagents.mdc).

  5. Find existing patterns to reuse. Prefer the codebase's existing helpers, conventions, and abstractions over inventing new ones. Check whether a dependency or utility already exists before adding one.

  6. Consult project docs/rules for the area you are touching. See project-adapter.mdc for which docs and rules apply.

Working memory

Fill the File ledger section of the task-memory file: for each relevant file, record path -> one-line summary + the line ranges that matter. This lets you avoid re-reading unchanged files later (drift is detected via git — see working-memory.mdc).

Read the full file on GitHub · 60 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. 4d ago First seen · 60 lines · 24 tokens per session scan A 6c7b8cd84812

Subscribe to this mod's changes

02-context-grounding is a cursor rule published in the GitHub repository AG-Dmitry/composer-enhanced (3 stars, last pushed 2mo ago), licensed MIT. It adds 24 tokens to every session and 577 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.