← All Docs

MCP Server

Tool Reference: PM

PM tools manage Webase projects, features, and project memory. They're independent from the App Gen tool family and let an agent run a Kanban-style workflow: list projects, claim the next feature, mark features in progress, complete them, and append implementation summaries.

Endpoint: https://www.webase.com/mcp/messages

Protocol: JSON-RPC over HTTP using MCP tools/list and tools/call.

Required scopes: pm_read for read tools, pm_write for write tools, or claudeai for both.

This surface exposes 14 tools. The list is generated from the live tool classes registered to this MCP server, so it always reflects what is actually callable in production.

append_implemented_summary WRITE

Use this after implementation work ships to append a concise summary of what was delivered to the project's implemented summary.

Input schema

project_id_or_slug Required string

Project identifier. Accepts either the numeric id or the friendly slug.

text Required string

Summary text to append to the implemented summary.

claim_next_feature WRITE Destructive

Use this in an automated implementation loop to atomically reserve the next ready, unblocked feature for a project and mark it running for the current agent.

Input schema

project_id_or_slug Required string

Project identifier. Accepts either the numeric id or the friendly slug.

agent_name string

Optional agent name to record on the claimed feature, such as codex.

complete_feature WRITE Destructive

Use this after implementation is verified to mark a feature done, record completion metadata, and clear it from the active execution queue.

Input schema

feature_id_or_slug Required string

Feature identifier. Accepts either the numeric id or the friendly slug.

agent_name string

Optional agent name to record on the completed feature, such as codex.

create_feature WRITE

Use this to create a new feature in a project after checking project context and existing features. Do not use it to update an existing feature.

Input schema

project_id_or_slug Required string

Project identifier. Accepts either the numeric id or the friendly slug.

title Required string

Short feature title.

description Required string

Detailed feature description for planning and implementation.

priority string

Optional priority. Allowed values: low, medium, high, urgent.

roadmap_bucket string

Optional roadmap bucket. Allowed values: now, next, later.

depends_on_feature_id null | integer

Optional dependency feature id in the same project. Use null to omit any dependency.

get_feature READ

Use this to fetch the full feature record when you need current implementation status, dependency state, or detailed feature content.

Input schema

feature_id_or_slug Required string

Feature identifier. Accepts either the numeric id or the friendly slug.

get_next_feature READ

Use this when an implementation agent needs the next ready, unblocked, non-running feature for a project. Returns the selected feature plus project context.

Input schema

project_id_or_slug Required string

Project identifier. Accepts either the numeric id or the friendly slug.

get_project READ

Use this to fetch the full project record when you need project metadata such as name, key, status, description, context, or implemented summary.

Input schema

project_id_or_slug Required string

Project identifier. Accepts either the numeric id or the friendly slug.

get_project_context READ

Use this when you need project memory for planning or implementation. Returns only the project context and implemented summary.

Input schema

project_id_or_slug Required string

Project identifier. Accepts either the numeric id or the friendly slug.

list_features READ

Use this to list all features for a specific project before reading one in detail, updating one, or proposing new work.

Input schema

project_id_or_slug Required string

Project identifier. Accepts either the numeric id or the friendly slug.

list_projects READ

Use this to list the current user's projects before selecting one to inspect, update, or create features in. Returns only a small project summary.

Input schema

This tool takes no arguments.

search_project_memory READ

Use this to search project memory for prior decisions, shipped work, implementation notes, and related features before proposing or creating new work.

Input schema

project_id_or_slug Required string

Project identifier. Accepts either the numeric id or the friendly slug.

query Required string

Plain-text search query used against project context, implemented summary, and feature title, description, implementation plan, implementation summary, and failure summary.

update_feature WRITE Destructive

Use this to update an existing feature's planning or execution fields such as title, description, implementation plan, implementation summary, failure summary, status, priority, roadmap bucket, or dependency.

Input schema

feature_id_or_slug Required string

Feature identifier. Accepts either the numeric id or the friendly slug.

fields_to_update Required object

Hash of feature fields to update. Provide only the fields that should change.

update_project WRITE Destructive

Use this to update mutable project metadata such as the project description.

Input schema

project_id_or_slug Required string

Project identifier. Accepts either the numeric id or the friendly slug.

fields_to_update Required object

Hash of project fields to update. Provide only the fields that should change.

update_project_context WRITE Destructive

Use this to replace the project's planning context with the latest authoritative context, constraints, or operating notes.

Input schema

project_id_or_slug Required string

Project identifier. Accepts either the numeric id or the friendly slug.

context Required string

Full replacement project context text.

Typical session

tools/call list_projects                          # find my projects
tools/call get_project { project_id_or_slug: "..." }
tools/call get_next_feature { project_id_or_slug: "..." }
tools/call claim_next_feature { project_id_or_slug: "..." }
# (do the work — possibly through the App Gen MCP surface or
#  outside Webase entirely)
tools/call append_implemented_summary { ... }
tools/call complete_feature { feature_id: ... }

Notes

PM tools are also exposed on the managed and external endpoints, so a single agent connection can run both project management and app generation.
The PM tool surface uses its own scope prefix (pm_read / pm_write) — see Authentication & Scopes.