mcp-tools
Skill16 starsChanged 5 days agoMITView on GitHub
MCP tool decision tree and MCP-first fallback strategy. TRIGGER when: choosing whether to use an MCP tool versus a built-in, or an MCP tool is available for a task. SKIP: semantic-memory MCP usage specifically (use semantic-memory-mcp).
3 other projects use a version of this.
Adopted by
Most recently taken up by
- sazardev/gocatheir own version, first seen Oct 8, 2026
- operately/operatelytheir own version, first seen Oct 8, 2026
- extra-org/extratheir own version, first seen Jun 8, 2026
Other versions3 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 3
- operately/operately★ 618same name
- sazardev/goca★ 299same name
- extra-org/extra★ 119same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository39 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- agent-commsSkill
- agent-memorySkill
- api-designSkill
- context-engineeringSkill
- database-optimizationSkill
- distillSkill
- docker-templatesSkill
- error-handlingSkill
- frontend-designSkill
- github-actions-templateSkill
- git-operationsSkill
- logging-standardsSkill
- monitoring-observabilitySkill
- mui-stylingSkill
- necessity-firstSkill
- pattern-recognitionSkill
- planning-methodologySkill
- python-patternsSkill
- quality-validationSkill
- react-patternsSkill
- research-methodologySkill
- reverse-engineer-anythingSkill
- security-review-checklistsSkill
- semantic-memory-mcpSkill
- skill-authoringSkill
- sofa-searchSkill
- spec-designSkill
- spec-developSkill
- spec-researchSkill
- spec-reviewSkill
- spec-workflowSkill
- state-managementSkill
- systematic-debuggingSkill
- testing-strategySkill
- ui-ux-pro-maxSkill
- verification-before-completionSkill
- watch-patternsSkill
- worktree-managementSkill
The file
We split this skill into the rules it gives an agent
2 rules. 0 are common, 1 is distinctive, and 1 is very specific to this project.
0
Common
Most of the 3 edited versions we read keep these.
1
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 12
No MCP tool matches? Use built-in tools (Bash, Grep, WebSearch, etc.).
- Very specific: line 62
Use search_context for documents and notes, specs, conversation transcripts and Gitea issues/PRs/commits.
“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
MCP Tools Decision Tree
Priority Order
- Does an MCP tool exist for this operation? Use it first.
- Did the MCP tool fail (auth missing, plugin unavailable)? Fall back to built-in.
- No MCP tool matches? Use built-in tools (Bash, Grep, WebSearch, etc.).
MCP Plugin Quick Reference
| Plugin | Transport | Key Tools | Agents |
| context7 | stdio | mcp__context7__resolve-library-id, mcp__context7__get-library-docs | researcher, developer |
| playwright | stdio | mcp__playwright__browser_navigate, mcp__playwright__browser_screenshot | developer, debugger |
| chrome-devtools | stdio | mcp__chrome-devtools__* | developer, debugger |
| eslint | stdio | mcp__eslint__* | developer, reviewer |
| sonarqube | docker | mcp__sonarqube__* | developer, reviewer |
| sequential-thinking | stdio | mcp__sequential-thinking__* | architect, debugger |
| postgres-mcp | stdio | mcp__postgres-mcp__* | developer, optimizer |
| redis-mcp | stdio | mcp__redis-mcp__* | developer, devops, debugger |
| docker | stdio | mcp__docker__* | devops |
| cron | stdio | mcp__cron__* | devops |
| ssh-mcp | stdio | mcp__ssh-mcp__* | devops |
| github | http | mcp__github__* | gitops, architect |
| google-sheets | stdio | mcp__google-sheets__* | researcher, tech-writer |
| slack | sse | mcp__slack__* | tech-writer |
| asana | sse | mcp__asana__* | architect |
| supabase | http | mcp__supabase__* | optimizer |
| firebase | stdio | mcp__firebase__* | devops, optimizer |
| memory | stdio | 13 tools across search/store/notes/ingest/session tiers (see below) | tiered — see Access Control |
Semantic Memory MCP
- Transport: stdio (Python,
venv/bin/python -m mcp_servers.semantic_memory) - Source: MCP servers in your backend directory (internal, built on
fastmcp) - Auth:
DATABASE_URL(PostgreSQL connection string),SEMANTIC_MEMORY_ENABLED=true - Config:
.mcp.jsonat project root
Tools (13 total)
| Tool | Purpose | Parameters |
search_context | Search ingested context chunks (Qdrant hybrid retrieval). Returns doc_id, source, kind, title, date, session_id, content, score (relative rank, not a probability) | query (required), corpus (homelab/platform/projects or a Gitea repo name), path (exact doc path in corpus), source_type (markdown/gitea), kind (note/spec/issue/pr/commit), session_id, top_k |
semantic_search | Search memories by similarity | query (required), project_id, agent_name, top_k, threshold |
semantic_recall | Recall relevant memories as markdown | context (required), agent_name, project_id, top_k |
semantic_store | Store a new memory with embedding | content (required), agent_name (required), project_id, conversation_id, task_id, tags, content_type |
store_note | Persist a note document | corpus, relative_path, content (see notes-tier agents) |
read_note | Read a note document | corpus, relative_path |
list_notes | List note documents | corpus |
trigger_ingest | Queue ONE document (corpus + relative_path) for immediate re-indexing so a note just written via store_note becomes searchable in seconds instead of up to 15 minutes. Path-scoped only — cannot trigger a full backfill. Returns once queued; does not wait for indexing to finish (~3s to become searchable). Only useful to agents that also have store_note — otherwise you'd be re-indexing someone else's file. | corpus (required), relative_path (required) |
list_sessions | List memory sessions (catalog) | — |
get_session | Get a session's details | session_id |
semantic_list | List stored memories (catalog) | project_id |
semantic_delete | Delete a memory | memory_id (granted to no agent) |
archive_session | Archive a session | session_id (granted to no agent) |
Use search_context for documents and notes, specs, conversation transcripts and Gitea issues/PRs/commits; use semantic_search for short distilled memories.
Access Control
| Tier | Tools | Agents |
| Write (search/recall/store) | search_context, semantic_search, semantic_recall, semantic_store | developer, architect, debugger, analyst, researcher, reviewer, optimizer, tech-writer |
| Read-only | search_context, semantic_search, semantic_recall | devops, gitops, mcp-builder, prompt-engineer |
| Notes | store_note, read_note, list_notes | architect, researcher, tech-writer |
| Ingest trigger | trigger_ingest | architect, researcher, tech-writer — narrow grant, restricted to the agents that also hold store_note; broader access would invite wasted LLM context-gen calls re-indexing files the caller didn't write |
| Session catalog | list_sessions, get_session, semantic_list | coordinator |
| Delete/archive | semantic_delete, archive_session | nobody |
corpus (top-level NFS share directory: homelab, platform, projects) and project_id (scaffold:<hash>, one per repo) are different axes — corpus scopes the notes/ingest tools above, project_id scopes the semantic_* (mem0) tools. Don't pass one where the other is expected.
The tools: frontmatter of each agent is authoritative for grants. For detailed usage guidance (when to search, when to store, quality gates), see the semantic-memory-mcp skill.
Fallback Strategy
| Operation | MCP Tool (Priority) | Fallback |
| Library docs lookup | context7 | WebSearch |
| UI verification | playwright | Manual browser check |
| Code quality | sonarqube | python3 devops/sonarqube.py via Bash |
| Linting | eslint | npx eslint via Bash |
| Database query | postgres-mcp | psql via Bash |
| Cache inspection | redis-mcp | redis-cli via Bash |
| Container ops | docker | docker via Bash |
| Git operations | github | gh CLI via Bash |
| Memory search | memory | File-based agent-memory |
| Remote server | ssh-mcp | ssh via Bash |
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.