Works with the AI Agents you already use
CallGraph QA — Agensi Skill Pack
Analyze, index, and query Python codebases using AST-aware chunking and call-graph enriched RAG.
$5
CallGraph QA — Agensi Skill Pack
Example session with this skill installed
Analyze the implementation of the checkout logic in this repo. Explain how it interacts with the payment gateway based on the call graph.
- Read your context and instructions
- Compiled the callgraph qa
I have analyzed the checkout.py and payments/gateway.py modules using AST parsing and graph-expanded retrieval. The semantic search identified the process_checkout function, and graph expansion successfully pulled in the dependent gateway calls.
Codebase Analysis: Checkout to Payment Flow
| Component | Symbol | Location | Role |
|---|---|---|---|
| Primary Match | process_checkout | services/checkout.py:45 | Main entry point for cart finalization. |
| Direct Callee | validate_session | auth/session.py:12 | Verifies user token before processing. |
| Direct Callee | charge_card | payments/gateway.py:88 | External API wrapper for Stripe/Braintree. |
| Caller | CheckoutAPIView.post | api/views.py:210 | REST endpoint triggering the service. |
Control Flow Summary
- The
CheckoutAPIViewreceives a POST request and callsprocess_checkout. process_checkoutinvokesvalidate_sessionto ensure the user is authenticated.- Upon successful validation, the service initializes the
PaymentGatewayand callscharge_card. - Static analysis shows
charge_cardis an async function requiring theamountandcurrencyattributes from theOrderobject.
Call Graph Context
The retrieval included charge_card because it is explicitly called within the body of process_checkout. While charge_card did not match the initial semantic query for "checkout logic" as strongly as the service layer, its structural relationship as a callee provides the necessary context to explain the payment integration.
Next steps
- Inspect
payments/gateway.pyfor specific error handling logic during API timeouts. - Review the
Orderclass definition to verify attribute types passed to the gateway. - Verify if the
charge_cardcall is wrapped in a retry decorator.
Connects securely to your tools. The creator never sees your data.
About this skill
Traditional RAG often fails on complex codebases because simple text chunks lose the structural context of how functions interact. This skill pack provides a specialized set of tools for structural Python analysis, graph-enriched retrieval, and evidence-based codebase Q&A.
What it does
- AST-Aware Analysis parses Python source code to identify functions and classes, ensuring code is never split mid-block.
- Structural Chunking maintains module paths, qualified symbol names, and line-range metadata for every code segment.
- Graph Expansion enriches standard vector search results by including immediate callers and callees in the retrieval context.
- Static Call Extraction walks function bodies to map relationships between symbols without requiring runtime execution.
- Evidence-Based Q&A answers technical questions about repository behavior using confirmed source citations and data flow tracing.
How it works
- Parse the Python codebase using the
astmodule to map top-level functions, async definitions, and class methods. - Index code chunks by preserving complete syntactical units and attaching symbol-level metadata.
- Query the index using vector similarity, then expand the context window by fetching one-hop graph neighbors (callers/callees).
- Synthesize an answer that distinguishes between direct semantic matches and structural graph matches.
Frameworks & tools
This tool uses Python's ast module for static analysis. It is designed for RAG systems using vector databases and LLMs, specifically targeting Python repositories with complex call hierarchies.
Why this beats prompting it yourself
Generic prompts treat code as raw text, leading to broken context and "lost" logic when functions span multiple chunks. This skill enforces structural integrity and graph-based relationships, ensuring the agent sees the code's execution path rather than just disconnected snippets.
Use cases
- Codebase Onboarding to map out where specific features are implemented in an unfamiliar repository.
- Root Cause Analysis to trace data flow from entry points to bug-prone utility functions.
- RAG System Design to architect code-aware retrieval pipelines that use structural graph context.
- Dependency Mapping to identify which functions will be impacted by a refactor or change.
Known limitations
Does not support full type inference or dynamic dispatch resolution. Limited to Python source code. Graph expansion is restricted to one-hop caller/callee relationships based on static function names.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for LLM Ops.
Supports Python only. Static analysis cannot resolve dynamic dispatch, monkey-patching, or decorators. Graph expansion is restricted to one-hop caller/callee relationships.
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 today
- Software
Python
Node.js