NakshAstraMCP: Agent Behavioral Guide (v3.10.1)

This repository is optimized for use with NakshAstraMCP. MANDATORY: Always prioritize MCP tools over manual file-dumps or generic search.

---

🛠️ MCP Capabilities (Ground Truth)

Tool	Priority	Usage Instruction
deep_context	Primary	Start here for all tasks. Locates files and adds their immediate graph neighbors (1-hop) for structural context.
search_codebase	Discovery	Broad “grep-style” keyword search when you lack a specific entry point.
find_symbol	Surgical	Precise lookup for classes/functions by name. Use for “Go to Definition.”
read_file	Verification	MANDATORY before any edit. Read the full file content to avoid assumptions.
find_references	Audit	Trace usages (Blast-radius) before changing shared or critical functions.
server_status	Diagnostic	Use only if indexing or memory issues are suspected.

---

🔄 Standardized Workflow

Follow this pattern for every non-trivial change:

1. deep_context: Discover relevant files and architectural neighbors.
2. find_symbol: Locate the exact definition (line ranges) for the target logic.
3. read_file: MUST read the full file content before proposing or drafting code.
4. find_references: Audit call sites to ensure no regressions are introduced.
5. search_codebase: Supplementary use only if deep_context misses specific keywords.

---

🛡️ Guardrails

• No Snippet Assumptions: Snippets are for discovery only. Never assume a full definition from a snippet.
• Verification First: Never edit a file you haven’t read fully via read_file.
• Tool Efficiency: Prefer deep_context for architectural discovery over scattered search_codebase calls.
• Security Protocols: All paths are jailed. If a tool fails with a security error, verify the path is within a registered workspace.

---

🤖 AI Self-Configuration Prompt

If you are an AI assistant tasked with maintaining or exploring this codebase, execution of the following sequence is required for initial synchronization. MANDATORY: You must use NakshAstraMCP tools specifically for this discovery; generic searches are prohibited.

Instruction for AI Agents:

• Run Below Prompt once when setting up new repository

“Perform an initial environment discovery by scanning the root directory for .md files and using deep_context on central modules to identify structural entry points. Create or update a local DOCUMENTATION_INVENTORY.md to track key architectural components and their respective guides for the user. CRITICAL: Upon successful completion, add a direct reference/link to the DOCUMENTATION_INVENTORY.md within the relevant documentation section of agent.md, then remove this ‘AI Self-Configuration Prompt’ section entirely, while strictly preserving all other behavioral guidelines and MCP instruction sets within the file.”

---

IMPORTANT INSTRUCTIONS (follow these strictly): • DO NOT write any greeting, introduction, or conclusion. Start with the answer immediately. No ‘Great question!’ or ‘I hope this helps!’ - if you do, you have FAILED.