MCP Resources¶

Context Teleport exposes 16 MCP resources that provide read-only access to the context store. Resources follow the context:// URI scheme and return JSON strings.

Resources are organized into seven groups: Project Metadata, Knowledge, Conventions, Decisions, State and History, Skills, and Activity.

Project Metadata¶

`context://manifest`¶

Project metadata and configuration. Returns the full Manifest model as JSON.

URI: context://manifest

Response schema:

{
  "schema_version": "0.4.0",
  "project": {
    "name": "my-project",
    "id": "a1b2c3d4-...",
    "repo_url": "https://github.com/org/my-project"
  },
  "adapters": {
    "claude_code": {"enabled": true}
  },
  "team": {
    "members": [
      {"name": "alice", "machine": "laptop", "added": "2025-05-01T10:00:00Z"}
    ]
  },
  "created_at": "2025-05-01T10:00:00Z",
  "updated_at": "2025-05-10T14:30:00Z"
}

`context://summary`¶

High-level project context summary. Returns an aggregated overview computed from the store state, including counts and key lists.

URI: context://summary

Response schema:

{
  "project": "my-project",
  "knowledge_count": 12,
  "convention_count": 3,
  "decisions_count": 5,
  "skills_count": 3,
  "current_task": "Implementing auth middleware",
  "blockers": ["Waiting for API spec"]
}

Knowledge¶

`context://knowledge`¶

List all knowledge entries with their keys, content, and scopes.

URI: context://knowledge

Response schema:

[
  {
    "key": "architecture",
    "content": "## Architecture\n\nMicroservices with gRPC...",
    "scope": "public"
  },
  {
    "key": "credentials-setup",
    "content": "...",
    "scope": "private"
  }
]

Note

All entries are returned regardless of scope. Scope filtering happens at the adapter export and onboarding prompt level, not at the resource level.

`context://knowledge/{key}`¶

Read a single knowledge entry by key. Returns the key and full markdown content.

URI: context://knowledge/{key}

URI parameters:

Parameter	Type	Description
`key`	`str`	Knowledge entry identifier

Response schema:

{
  "key": "architecture",
  "content": "## Architecture\n\nMicroservices with gRPC..."
}

Error response:

{"error": "Knowledge entry 'nonexistent' not found"}

Conventions¶

`context://conventions`¶

List all team conventions with their keys, content, and scopes.

URI: context://conventions

Response schema:

[
  {
    "key": "git",
    "content": "Always use feature branches.\nCommit early, commit often.",
    "scope": "public"
  },
  {
    "key": "environment",
    "content": "No sudo. Use venvs. Docker when needed.",
    "scope": "public"
  }
]

`context://conventions/{key}`¶

Read a single convention by key. Returns the key and full markdown content.

URI: context://conventions/{key}

URI parameters:

Parameter	Type	Description
`key`	`str`	Convention identifier

Response schema:

{
  "key": "git",
  "content": "Always use feature branches.\nCommit early, commit often."
}

Error response:

{"error": "Convention 'nonexistent' not found"}

Decisions¶

`context://decisions`¶

List all architectural decisions with their IDs, titles, statuses, and scopes.

URI: context://decisions

Response schema:

[
  {
    "id": 1,
    "title": "Use PostgreSQL for persistence",
    "status": "accepted",
    "scope": "public"
  },
  {
    "id": 2,
    "title": "Adopt gRPC for service communication",
    "status": "proposed",
    "scope": "public"
  }
]

Status values: proposed, accepted, deprecated, superseded

`context://decisions/{id}`¶

Read a single decision by ID or title. Returns the full ADR content.

URI: context://decisions/{id}

URI parameters:

Parameter	Type	Description
`id`	`str`	Decision ID (numeric) or title

Response schema:

{
  "id": 1,
  "title": "Use PostgreSQL for persistence",
  "status": "accepted",
  "context": "Need ACID transactions and JSON support",
  "decision": "PostgreSQL 16 with pgvector extension",
  "consequences": "Team needs PostgreSQL expertise"
}

Error response:

{"error": "Decision '99' not found"}

State and History¶

`context://state`¶

Current active session state. Returns the ActiveState model as JSON.

URI: context://state

Response schema:

{
  "current_task": "Implementing auth middleware",
  "blockers": ["Waiting for API spec"],
  "progress": {"auth": "50%", "tests": "done"},
  "last_agent": "claude-code",
  "last_machine": "laptop",
  "updated_at": "2025-05-10T14:30:00Z"
}

`context://history`¶

Recent session history. Returns a list of SessionSummary objects ordered by start time.

URI: context://history

Response schema:

[
  {
    "id": "a1b2c3d4-...",
    "agent": "claude-code",
    "user": "",
    "machine": "",
    "started": "2025-05-10T14:00:00Z",
    "ended": null,
    "summary": "Set up database schema and migrations",
    "knowledge_added": ["tech-stack", "db-schema"],
    "decisions_added": ["3"],
    "skills_used": ["run-tests"]
  }
]

Skills¶

`context://skills`¶

List all agent skills with their names, descriptions, and scopes.

URI: context://skills

Response schema:

[
  {
    "name": "deploy-staging",
    "description": "Deploy the application to staging environment",
    "scope": "public"
  },
  {
    "name": "run-tests",
    "description": "Execute the full test suite with coverage",
    "scope": "public"
  }
]

`context://skills/{name}`¶

Read the full SKILL.md content for a specific skill, including YAML frontmatter and the markdown body.

URI: context://skills/{name}

URI parameters:

Parameter	Type	Description
`name`	`str`	Skill name

Response schema:

{
  "name": "deploy-staging",
  "description": "Deploy the application to staging environment",
  "content": "---\nname: deploy-staging\ndescription: Deploy the application to staging environment\n---\n\n## Steps\n\n1. Build the Docker image\n2. Push to registry\n3. Update Kubernetes deployment"
}

Error response:

{"error": "Skill 'nonexistent' not found"}

`context://skills/stats`¶

Aggregated usage and feedback statistics for all skills. Returns computed SkillStats objects.

URI: context://skills/stats

Response schema:

[
  {
    "skill_name": "deploy-staging",
    "usage_count": 15,
    "avg_rating": 4.2,
    "rating_count": 5,
    "last_used": "2025-05-10T14:30:00Z",
    "needs_attention": false
  },
  {
    "skill_name": "run-tests",
    "usage_count": 42,
    "avg_rating": 2.5,
    "rating_count": 4,
    "last_used": "2025-05-10T12:00:00Z",
    "needs_attention": true
  }
]

Attention flag

needs_attention is true when a skill has 2 or more ratings and the average rating is below 3.0.

`context://skills/{name}/feedback`¶

All feedback entries for a specific skill. Returns the raw SkillFeedback records from the .feedback.ndjson sidecar file.

URI: context://skills/{name}/feedback

URI parameters:

Parameter	Type	Description
`name`	`str`	Skill name

Response schema:

[
  {
    "id": "e5f6g7h8-...",
    "agent": "claude-code",
    "rating": 4,
    "comment": "Clear instructions, easy to follow",
    "timestamp": "2025-05-10T14:30:00Z"
  },
  {
    "id": "i9j0k1l2-...",
    "agent": "cursor",
    "rating": 2,
    "comment": "Missing rollback step",
    "timestamp": "2025-05-09T10:00:00Z"
  }
]

Error response:

{"error": "Skill 'nonexistent' not found"}

`context://skills/{name}/proposals`¶

All improvement proposals for a specific skill. Returns SkillProposal records from the .proposals/ sidecar directory.

URI: context://skills/{name}/proposals

URI parameters:

Parameter	Type	Description
`name`	`str`	Skill name

Response schema:

[
  {
    "id": "m3n4o5p6-...",
    "skill_name": "deploy-staging",
    "agent": "claude-code",
    "rationale": "Add rollback step for failed deployments",
    "proposed_content": "---\nname: deploy-staging\n...",
    "diff_summary": "@@ -10,3 +10,7 @@\n ...",
    "status": "pending",
    "created_at": "2025-05-10T14:30:00Z",
    "resolved_at": null,
    "resolved_by": ""
  }
]

Proposal status values: pending, accepted, rejected, upstream

Error response:

{"error": "Skill 'nonexistent' not found"}

Activity¶

`context://activity`¶

List all team activity entries with stale indicators. Returns all entries regardless of staleness; stale entries are flagged for the consumer to filter or display as needed.

URI: context://activity

Response schema:

[
  {
    "member": "alice",
    "agent": "claude-code",
    "machine": "laptop",
    "task": "Fixing DRC violations",
    "issue_ref": "#42",
    "status": "active",
    "updated_at": "2025-05-10T14:30:00+00:00",
    "stale": false
  }
]

Stale threshold

Entries older than 48 hours are marked stale: true. The context_onboarding prompt omits stale entries from its output; the resource always returns the full list.

Resource Access Patterns¶

Resources are designed for two primary use cases:

Onboarding -- An agent reads context://summary, context://conventions, context://knowledge, context://decisions, and context://skills to understand the project state before starting work. The context_onboarding prompt automates this.
Targeted lookup -- An agent reads a specific entry (context://knowledge/{key}, context://conventions/{key}, context://decisions/{id}, context://skills/{name}) when it needs detailed content about a particular topic.

Resources are read-only

Resources provide read access only. To modify context, use the corresponding MCP tools. For example, read via context://knowledge/{key} and write via context_add_knowledge().

MCP Resources¶

Project Metadata¶

context://manifest¶

context://summary¶

Knowledge¶

context://knowledge¶

context://knowledge/{key}¶

Conventions¶

context://conventions¶

context://conventions/{key}¶

Decisions¶

context://decisions¶

context://decisions/{id}¶

State and History¶

context://state¶

context://history¶

Skills¶

context://skills¶

context://skills/{name}¶

context://skills/stats¶

context://skills/{name}/feedback¶

context://skills/{name}/proposals¶

Activity¶

context://activity¶

Resource Access Patterns¶

`context://manifest`¶

`context://summary`¶

`context://knowledge`¶

`context://knowledge/{key}`¶

`context://conventions`¶

`context://conventions/{key}`¶

`context://decisions`¶

`context://decisions/{id}`¶

`context://state`¶

`context://history`¶

`context://skills`¶

`context://skills/{name}`¶

`context://skills/stats`¶

`context://skills/{name}/feedback`¶

`context://skills/{name}/proposals`¶

`context://activity`¶