Skip to main content
Glama
oleksiijko

io.github.oleksiijko/pmb-ai

by oleksiijko

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault

No arguments

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

CapabilityDetails
tools
{
  "listChanged": true
}
logging
{}
prompts
{
  "listChanged": false
}
resources
{
  "subscribe": false,
  "listChanged": false
}
extensions
{
  "io.modelcontextprotocol/ui": {}
}
experimental
{}

Tools

Functions exposed to the LLM to take actions

NameDescription
recallA

Search memory for anything about the user/past/project. recall(query, top_k=5). Returns results + auto-attached lessons (read & FOLLOW them) + project_context. Trust results with score>0.2 as the user's recorded reality.

project_overviewA

One-call full context for a NAMED project at the start of work. project_overview(name) → lessons (rules to follow), decisions, open goals, recent activity, related entities.

prepareA

READ-FIRST bundle at the start of work on a known project. prepare(message=<the user's message>) returns project_context, surfaced lessons (each with surface_id - FOLLOW them, then mark_lesson_followed), recent_activity and open_goals in one ~10ms call. Replaces several recall() calls.

record_keyed_factA

Upsert a mutable personal attribute. record_keyed_fact(subject, attribute, value) - e.g. user/city/Tampa. A new value SUPERSEDES the old under one canonical key instead of piling up.

find_lessonsA

Standalone 'what procedural rules apply to X'. find_lessons(query, project?) → lessons with surface_id; project scope excludes explicit lessons from other projects while retaining generic rules. FOLLOW them, then mark_lesson_followed.

mark_lesson_followedA

Report whether a surfaced lesson changed your behaviour. mark_lesson_followed(surface_id, followed=True|False, note='...', applicable=True|False). Use applicable=False when the lesson was irrelevant, not followed=False. Call after acting on a lesson - powers the self-improvement loop.

session_briefA

Re-orient after YOUR context compacted. session_brief() → what THIS session decided/built. Don't re-ask the user.

record_batchA

⚡ PREFERRED for any message with multiple memories - stores N atomic items in ONE call (each ~3-5s of agent thinking saved vs separate record_* calls). items: list of dicts, each with a type: fact{content,importance} | fact_tree{main,subfacts[],importance} | lesson{content,project?} | goal{title,status,due_at} | plan{title} (future intent) | activity{content,kind} | milestone{chain_name,title,state}. ONE record_batch per turn; use ABSOLUTE dates.

update_goalA

Move a goal's status/progress. update_goal(goal_ulid, status='in_progress'|'done', progress=0-100, note='...'). Records a goal_update event so the goal's history is preserved.

list_goalsA

List open goals. list_goals(status='in_progress'). For 'what are my goals/what's in flight'.

Prompts

Interactive templates invoked by user choice

NameDescription

No prompts

Resources

Contextual data attached and managed by the client

NameDescription

No resources

Latest Blog Posts

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/oleksiijko/pmb'

If you have feedback or need assistance with the MCP directory API, please join our Discord server