Two tools with basic schemas and descriptions, but significant gaps in parameter documentation and output structure. search_vulnerability has reasonable descriptions but lacks output schema documentation. search_techniques description is verbose (>200 chars) and mentions implementation details (ChromaDB, 5-15 min indexing) that belong in docs, not the tool description. Neither tool documents return structure, pagination, or error recovery paths. Parameter descriptions are present but lack constraints (e.g., max_results range, source enum values). No tool annotations (readOnlyHint, idempotentHint). Error handling is absent, no guidance on what to do if NVD/CISA/GitHub APIs fail or timeout.

• Document output schema for search_vulnerability: return a structured object with 'nvd_results' (array of {cve_id, cvss_score, description}), 'cisa_kev_results' (array of {cve_id, vendor, product, due_date}), 'github_poc_results' (array of {repo_name, stars, url, description}), and 'errors' (array of {source, error_message}).
• Rewrite search_techniques description to 50 - 100 chars: 'Search HackTricks and PayloadsAllTheThings for pentesting techniques using natural language queries.' Move indexing details to server docs.
• Add parameter constraints: max_results (1 - 50, default 5), top_k (1 - 50, default 5), source (enum: 'hacktricks' | 'payloads' | null, default null).
• Add tool annotations: @mcp.tool(readOnlyHint=True, idempotentHint=True) to both tools so LLMs know they are safe to retry.
• Enhance error handling: return structured error responses with actionable guidance, e.g., 'NVD API timeout. Try again in 30 seconds or use CISA KEV as fallback.' instead of raw exception strings.