- Home
- Skills
- Agents & Orchestration
- π Reference Grounding Checker
More screenshots
Works with the AI Agents you already use
π Reference Grounding Checker
It eliminates "hallucinated" refactoring plans before they break your build.
$15
π Reference Grounding Checker
Example session with this skill installed
I've drafted a plan to refactor the auth logic. Check this plan against the repo to make sure I haven't referenced any nonexistent functions or files.
- Read your context and instructions
- Compiled the reference grounding checker
Grounding Report
- Files Checked: 4
- References Verified: 12
- Findings:
- [High]
utils/token-helper.tsnot found. (Line 12) - [Med] Function
verifySession()is exported fromauth/core.ts, notauth/session.tsas suggested.
- [High]
- Suggestion: Update plan lines 12 and 15.
Connects securely to your tools. The creator never sees your data.
About this skill
What it does
The Reference Grounding Checker acts as a pre-execution safety layer for AI agents. It scans proposed implementation plans to verify that every file path, function name, import statement, and environment variable mentioned actually exists within your local codebase. It eliminates "hallucinated" refactoring plans before they break your build.
Why use this skill
Even advanced LLMs frequently reference non-existent utility functions or assume the presence of configuration variables that aren't there. This skill provides a deterministic check against the ground truth of your file system and source code. It is significantly more reliable than standard prompting because it uses a specialized static analysis scanner to perform hard lookups, rather than relying on the model's memory of the repo.
Supported tools
- Languages: Python, JavaScript, TypeScript.
- Environments: Local repositories, CI/CD pipelines, and agentic workflows like Claude Code or Cursor.
- Inputs: Markdown plans, text files, or direct chat history.
Expected Output
You receive a structured report detailing every reference checked, categorized findings by severity (e.g., Missing File, Unresolved Import), and specific remediation snippets to fix the plan's logic.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Agents & Orchestration.
Extraction is heuristic and works best on backtick-wrapped references, so dynamic or generated names can be missed or over-flagged. It resolves Python and JS/TS; other languages are out of scope for symbol and import resolution.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 19 days ago
Permissions
File Scopes
Notes: Read-only. Extracts referenced tokens from the plan and resolves them against the repo using the standard-library ast module. Reads no environment variables and writes nothing.