batch_scan

Instructions

Implementation Reference

• src/mcp/server.ts:319-384 (handler)

  The main handler for the batch_scan MCP tool. Accepts an array of up to 10 token mint addresses. For each mint, runs 5 analyses in parallel (token safety, honeypot check, holder analysis, Birdeye enrichment, and wallet reputation). Aggregates results using Promise.allSettled, computes a weighted final risk score, and returns verdicts (SAFE/CAUTION/HIGH_RISK/CRITICAL) for each token.

  // ── Tool: batch_scan ──────────────────────────────────────────────────
  server.tool(
      'batch_scan',
      `Scan multiple Solana tokens in a single call for portfolio-level risk assessment. Runs full_token_scan on each mint in parallel (max 10 per batch). Returns an array of results with verdicts for each token. Use this when evaluating a portfolio, watchlist, or multiple tokens from a pool discovery.`,
      {
          mints: z.array(z.string()).max(10).describe('Array of Solana token mint addresses to scan (max 10)'),
      },
      async ({ mints }) => {
          try {
              const results = await Promise.allSettled(
                  mints.map(async (mint) => {
                      const [safety, honeypot, holders, birdeye, walletIntel] = await Promise.all([
                          analyzeTokenSafety(connection, mint, undefined, false),
                          checkHoneypot(mint),
                          analyzeHolders(connection, mint),
                          enrichWithBirdeye(mint, BIRDEYE_API_KEY),
                          enrichCreatorReputation(mint, HELIUS_API_KEY),
                      ]);
  
                      let onChainScore = safety.riskScore;
                      if (honeypot.isHoneypot) onChainScore = Math.min(onChainScore + 30, 100);
                      if (holders.concentrated) onChainScore = Math.min(onChainScore + 15, 100);
  
                      const reputationScore = walletIntel.reputation?.riskScore ?? 0;
                      const marketScore = birdeye.marketRisk.score;
                      const finalScore = Math.round(
                          onChainScore * 0.60 + marketScore * 0.25 + reputationScore * 0.15
                      );
                      const verdict = finalScore === 0 ? 'SAFE'
                          : finalScore <= 15 ? 'CAUTION'
                          : finalScore <= 50 ? 'HIGH_RISK'
                          : 'CRITICAL';
  
                      return {
                          mint,
                          safe: safety.safe && !honeypot.isHoneypot && !holders.concentrated && marketScore < 30,
                          finalScore,
                          verdict,
                          honeypot: honeypot.isHoneypot,
                          liquidity: birdeye.overview?.liquidity ?? null,
                          marketFlags: birdeye.marketRisk.flags,
                          walletAge: walletIntel.reputation?.creatorAge ?? null,
                      };
                  })
              );
  
              const output = results.map((r, i) => {
                  if (r.status === 'fulfilled') return r.value;
                  return { mint: mints[i], error: r.reason?.message ?? String(r.reason) };
              });
  
              return {
                  content: [{
                      type: 'text' as const,
                      text: JSON.stringify({ scanned: output.length, results: output }, null, 2),
                  }],
              };
          } catch (e: unknown) {
              const msg = e instanceof Error ? e.message : String(e);
              return {
                  content: [{ type: 'text' as const, text: JSON.stringify({ error: msg }) }],
                  isError: true,
              };
          }
      },
  );

• src/mcp/server.ts:387-387 (registration)

  The batch_scan tool is registered via server.tool('batch_scan', ...) at line 320 of src/mcp/server.ts. The description tells the LLM this is for portfolio-level risk assessment of multiple tokens in a single call.

  }

• src/api/server.ts:128-128 (registration)

  The batch_scan tool is also listed in the MCP server card at /.well-known/mcp/server-card.json for Smithery/registry discovery, describing it as 'Scan up to 10 tokens in parallel for portfolio-level risk assessment.'

  { name: 'batch_scan', description: 'Scan up to 10 tokens in parallel for portfolio-level risk assessment.' },

• src/mcp/server.ts:323-325 (schema)

  Input schema for batch_scan: takes a 'mints' field which is a Zod array of strings with a maximum of 10 items. Validated using zod.

  {
      mints: z.array(z.string()).max(10).describe('Array of Solana token mint addresses to scan (max 10)'),
  },