LiuHe
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| MALONG_PORT | No | TCP port for the parse daemon (Windows). | 31001 |
| MALONG_SOCKET | No | Unix socket path to the parse daemon (Linux / macOS). | /tmp/malong-parse-$(id -u).sock |
| MALONG_PARSE_BIN | No | Binary used when the client auto-starts the daemon. | npm platform pkg / ~/.local/bin |
| MALONG_STATE_DIR | No | Where usage / feedback / edit-stats files are written. Override to redirect state (tests, sandboxed hosts). | ~/.config/malong |
| MALONG_PARSE_MODE | No | Parse transport. Only 'rust-service' is supported. | rust-service |
| MALONG_WS_GC_DAYS | No | Days a workspace index cache may sit untouched before health cleanup prunes it; 0 disables. | 14 |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| active_todos | Scan TODO/FIXME/XXX/HACK, prioritize by current work. Pure regex, no index. Pass current_files to boost. |
| edit_batch | MUST USE for all file edits. PREFERRED single-file edit path. Read first, then edit. Multiple edits, atomic, dry_run previews diff. Rollback/multi-file? edit_transaction. |
| call_chain | Callers/callees at a line; use when you know line, not symbol name. Symbol-level? impact_analysis. |
| code_quality | 5-dim shape-level quality probe (techDebt/arch/blast): scores, not findings. |
| code_review | Shape-level code check (deterministic regex, zero LLM): naming, comment coverage, long functions, duplicated blocks, diff review (SEARCH/REPLACE). Clears shallow noise before deep review — not a correctness verdict. |
| code_search | NL intent search when symbol name unknown (known name? symbol_search). DETERMINISTIC token+intent matching, NOT semantic/embedding search — long natural-language sentences degrade to loose file-level hits (0.5); prefer precise symbol/file names. |
| config_drift | Detect env vars/DB tables/services in code missing from .env.example/docker-compose. Pairs with exception_guard. |
| sweep_dead_code | Detect dead code: unused imports/functions, orphan files, unused_guard signal. Text-reference fallback counts CLI-string refs as alive (miss-over-delete bias). Remove via edit_transaction. Out of scope: dynamic/reflection wiring. |
| debug_runner | Run command/script/test and analyze failures: 14 error types, 4-language stack traces, suggested actions. script mode auto-picks runtime (js/py/go/rs/sh). Deterministic, zero LLM. |
| dep_graph | Import dependency graph of a file (transitive N levels) or the whole project (scope=project → nodes/edges/cycles). Circular deps? fix_imports. Changing a dep? impact_analysis first. |
| dependency_gatekeeper | Imports vs manifests: undeclared deps + install hints (Py/JS/Go/Rust). Pairs with guard_patterns. |
| diff_facts | After edit_transaction: AST symbol changes + caller/test sync. Stale tests? test_bridge. |
| edit_collision_guard | Two-step guard: record_read (snapshot) after reading, check before editing. Detects external edits. Pairs edit_transaction. |
| sandbox_validate | Pre-validate edits: pass file + new_content (full new text), dry-run syntax check. Then edit_transaction. |
| edit_transaction | DEFAULT multi-file edit path with rollback/preview/undo (begin→edit→commit). Use when unsure about changes. Single-file? edit_batch. For string params containing double quotes/backslashes, keep JSON escaping consistent across the whole string (escape uniformly or prefer full-width quotes), otherwise client-side parameter parsing fails. |
| exception_guard | Check raise/throw uses project exceptions vs builtins, module-scoped (falls back to project scope). Fix → verify test_bridge. Pairs config_drift. |
| feedback | Report malong tool issues/errors/ideas; collected locally. |
| find_tests | Find test files for a source file (naming + import reverse lookup). Run them? test_bridge. AUTO-FRESH: files modified after last index are re-indexed automatically on read — explicit reindex only needed for brand-new workspaces/files. |
| fix_imports | Detect unused imports, undefined symbols, circular deps. auto_fix applies. |
| gc | Trigger manual GC (needs --expose-gc). Use when memory is high. |
| git_worktree | Multi-file changes on isolated git branch: zero working-tree pollution, auto commit + ff-merge, verify_cmd. Rollback on failure. |
| guard_patterns | AST rules: no bare-except/debugger/eval. Custom via .ai-patterns.json. Pre-commit gate with dependency_gatekeeper. |
| health | System check + self-healing: DB integrity, tools, memory, semaphore. actions: check/restart/cleanup. |
| impact_analysis | BEFORE modifying/renaming: callers (blast radius), callees, risk (caller-count blast). Line-level? call_chain. Simple edit? read_symbol+write_symbol. AUTO-FRESH: files modified after last index are re-indexed automatically on read — explicit reindex only needed for brand-new workspaces/files. |
| inspect | Outline + references + call chain in one call (replaces read_outline/references/call_chain). |
| mock_sync | Detect mock/patch mismatches after signature changes. Pairs rename_symbol. |
| naming_consistency | New symbol names vs project style (snake/camel/Pascal) + verb consistency. Index required. |
| read_outline | File structure: functions/classes/signatures, no full read. depth=hierarchy nesting, NOT call depth. Known symbol? read_symbol. |
| patch_parser | Parse SEARCH/REPLACE patch text (text=...), dry-run apply. For external patch blocks. |
| read_symbol | MUST READ BEFORE WRITE. Symbol body + version (base_version for writes). Resolves by symbol_id or name+file. Full file? read_outline. AUTO-FRESH: files modified after last index are re-indexed automatically on read — explicit reindex only needed for brand-new workspaces/files. |
| references | Cross-file usages of a symbol, incl. usage counts. Use full member path (obj.method). Prefer over grep. AUTO-FRESH: files modified after last index are re-indexed automatically on read — explicit reindex only needed for brand-new workspaces/files. |
| reindex | MUST run FIRST: index workspace before ALL other malong tools. Over-threshold workspaces require two-step confirm (call → get token → confirm). blocking/force options. malong = LLM-native toolset. |
| rename_symbol | Execute a cross-file symbol rename (edit_transaction-based). Word-boundary, string/comment-aware. dry_run previews. To see what a rename would affect FIRST, use impact_analysis. |
| repo_map | Project code map: files ↔ symbols. For 100+ file codebases. focused+relevantEntities for token efficiency. |
| security_review | Scan for security patterns: injection (eval/exec/spawn/SQL/template), XSS, secrets, CORS*, insecure compare. Severity-graded, zero LLM. Suppress via malong-ignore or .ai-patterns.json. Out of scope: SSRF/XXE/deserialization/auth. |
| spec_gen | Generate module/API spec from a source file's symbols. |
| style_sniffer | Sniff project code style from samples → PROJECT_RULES.md content (returns text; output dir optional). |
| symbol_search | Find symbol definitions by name (substring). Usages? references. Before edit? impact_analysis. AUTO-FRESH: files modified after last index are re-indexed automatically on read — explicit reindex only needed for brand-new workspaces/files. |
| test_bridge | Run tests, parse output, enrich failures with context. actions: run/suggest/discover. Find tests? find_tests. |
| trace_symbol | Trace constant values + hardcoded copies (magic numbers). Function impact? impact_analysis. Rename? rename_symbol. |
| tsc_check | TS type-check via tsc --noEmit. |
| verify_pipeline | Run lint/test/typecheck from package.json. WARNING: MCP requests have a 120s hard timeout; default is 30s per stage (3 stages = 90s). Large test suites will time out — prefer stages=lint or test_bridge(action=run) for full test runs. |
| write_symbol | Guarded write path (one symbol) — vs edit_batch: index-aware with version/lock/undo, for symbol-level refactors. Read first (read_symbol) for base_version. For string params containing double quotes/backslashes, keep JSON escaping consistent across the whole string (escape uniformly or prefer full-width quotes), otherwise client-side parameter parsing fails. |
| write_symbols | Guarded write path (batch). All-or-nothing across files, deadlock-safe. Multi-file refactor? Use instead of N×write_symbol. New symbols: reindex first. For string params containing double quotes/backslashes, keep JSON escaping consistent across the whole string (escape uniformly or prefer full-width quotes), otherwise client-side parameter parsing fails. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/wulun811/LiuHe'
If you have feedback or need assistance with the MCP directory API, please join our Discord server