Mako Code Intel

Use this skill for static code queries that are best answered by syntax-aware search or diagnostics. It is not the general repo-orientation skill; use repo_map through /mako-ai:mako-discovery for that.

Tools

ast_find_pattern

Use for structural code pattern search over indexed TypeScript, TSX, JavaScript, and JSX files.

• Best for finding syntax shapes that text search may miss.
• Use for patterns such as function calls, JSX attributes, hook usage, imports, conditionals, or unsafe constructs.
• Keep searches bounded with focused patterns and file scopes when possible.
• TSX/JSX snippets starting with {, [, or < also run with parser context, so callers usually do not need to wrap lone object or JSX patterns. Check patternAttempts and match patternVariant when you need to know whether the original or auto-anchored form matched.
• List metavariables in captures when you need captured values returned.
• If the result is a false positive that should not recur, use finding_ack from /mako-ai:mako-workflow with the returned ackableFingerprint.

lint_files

Use for focused lint/static diagnostics against selected files.

• Best after editing files or when reviewing a known suspect area.
• Defaults to compact output; pass verbosity: "full" or maxFindings when broad diagnostics are actually needed.
• YAML rule packs under .mako/rules hot-reload in long-running MCP sessions.
• For helper-bypass checks, rule packs may declare canonicalHelper with a helper symbol and optional path; matching files that already reference the helper are suppressed, while bypasses emit producer/consumer context.
• Use the returned finding identity/fingerprint when acknowledging reviewed diagnostics.
• Do not use agent_feedback to suppress a diagnostic; use finding_ack.

extract_rule_template

Use after a fix lands and the removed shape should become future project knowledge.

• Pass the local git fixCommit; optionally pass baseCommit, filePath, and ruleIdPrefix.
• Treat the returned draftYaml as a draft. Review broad patterns and caveats before placing it under .mako/rules.
• Follow with rule_pack_validate and a focused lint_files run before relying on the new rule.

Feedback Logging

Log agent_feedback when a tool run here was notably useful, partial, noisy, stale, wrong, or wasted the turn. Skip routine calls.

Required procedure (see /mako-ai:mako-guide for full rules and reason-code vocabulary):

1. Call recall_tool_runs to get the prior run's requestId. Do not fabricate one — if no run is recalled, skip feedback.
2. Call agent_feedback with referencedToolName, referencedRequestId, grade: "full" | "partial" | "no", reasonCodes from the starter vocabulary in /mako-ai:mako-guide, and a short reason.

For individual false-positive or reviewed-accepted static findings from ast_find_pattern or lint_files, use finding_ack with the returned ackableFingerprint (ast) or identity.matchBasedId (lint) — never agent_feedback to suppress static findings. Rate the tool run's usefulness with agent_feedback; mark individual findings with finding_ack. See /mako-ai:mako-workflow for finding_ack usage.

See Also

• Use /mako-ai:mako-discovery for repo orientation and broad cross-surface search.
• Use /mako-ai:mako-graph for import impact, cycles, hotspots, and symbols.
• Use /mako-ai:mako-workflow for finding_ack and finding ack reports.