@dollhousemcp/mcp-server
Installed, started with no arguments, no environment and no input, and listed its tools. Run 2026-11-16.
Needs to start: nothing
Contacts while installing and starting: nothing
What it contacted
Nothing besides the npm registry it was installed from. Watched from outside its sandbox while it was installed and started with no input. Its tools may still contact things when called.
Starting it
npx -y -p @dollhousemcp/mcp-server@2.0.43 mcp-server
Installing a package runs its code. Use a sandbox for anything you have not read.
What its tools tell an agent
In the tool descriptions, nothing that talks about the user, other tools or private files, or uses hidden characters.
5 tools with commands in capitals or very long text
- Gives the agent orders in
mcp_aql_create:ful assistant", instructions: "You ARE a helpful assistant. ALWAYS provide clear, accurate responses." } } { operation: "create_element", element_type: "agent", param
- Very long description in
mcp_aql_create:3000 characters or more
- Very long description in
mcp_aql_read:3000 characters or more
- Gives the agent orders in
mcp_aql_execute:t execution from a handoff block with integrity validation IMPORTANT: Execute operations are potentially destructive (agents can perform any action) and non-idempotent
- Very long description in
mcp_aql_execute:3000 characters or more
Common and usually harmless: it is how many makers stress how a tool should be used.
Found by simple pattern rules in each tool's description (up to 3,000 characters), quoted as written. Parameter descriptions are not scanned. A clean result is not a safety review.
Tools
mcp_aql_create operation, element_type, params, operations
Additive, non-destructive operations.
Supported operations: create_element, import_element, addEntry, verify_challenge, release_deadlock, beetlejuice_beetlejuice_beetlejuice, record_execution_step, install_collection_content, submit_collection_content, init_portfolio, sync_portfolio, portfolio_element_manager, setup_github_auth, configure_oauth, import_persona
Element types: persona, skill, template, agent, memory, ensemble
These operations add new data without removing or overwriting existing content.
Quick start examples:
{ operation: "create_element", element_type: "persona", params: { element_name: "MyPersona", description: "A helpful assistant", instructions: "You ARE a helpful assistant. ALWAYS provide clear, accurate responses." } }
{ operation: "create_element", element_type: "agent", params: { element_name: "MyAgent", description: "Task executor", instructions: "Execute goals methodically. Report progress at each step.", goal: { template: "Complete: {objective}", parameters: [{ name: "objective", type: "string", required: true }] } } }
{ operation: "create_element", element_type: "memory", params: { element_name: "session-notes", description: "Session context and notes" } }
{ operation: "create_element", element_type: "ensemble", params: { element_name: "my-ensemble", description: "Combined element set", metadata: { elements: [{ element_name: "expert", element_type: "persona", role: "primary" }, { element_name: "analysis", element_type: "skill", role: "support" }] } } }
Valid ensemble roles: primary, support, override, monitor, core
{ operation: "addEntry", params: { element_name: "session-notes", content: "Remember this fact", tags: ["important"] } }
Note: addEntry content supports markdown (headers, lists, bold, tables, code blocks). Ensure markdown content is properly JSON-escaped — use \n for newlines, \" for quotes, and \\ for backslashes within the JSON string value.
Execution lifecycle — record agent progress (appends step records, like addEntry):
{ operation: "record_execution_step", params: { element_name: "code-reviewer", stepDescription: "Analyzed files", outcome: "success", findings: "Found 3 issues" } }
This is the normal next lifecycle call after mcp_aql_execute { operation: "execute_agent", ... }.
Response flow: record_execution_step returns { autonomy: { continue, factors, notifications? } }. Check autonomy.continue to decide whether to proceed. Check autonomy.notifications for permission_pending (gatekeeper blocks), autonomy_pause, or danger_zone alerts to relay to human operators.
Import & portfolio:
{ operation: "import_element", element_type: "skill", params: { element_name: "code-formatter", data: "..." } }
{ operation: "import_persona", params: { source: "/path/to/persona.md" } }
{ operation: "install_collection_content", params: { element_type: "persona", element_name: "Creative-Writer" } }
{ operation: "submit_collection_content", params: { element_type: "skill", element_name: "code-formatter" } }
{ operatiomcp_aql_read operation, element_type, params, operations
Safe, read-only operations.
Supported operations: activate_element, open_portfolio_browser, open_logs, open_metrics, open_permissions, open_setup, permission_prompt, evaluate_permission, get_effective_cli_policies, get_permission_authority, get_pending_cli_approvals, search, list_elements, get_element, get_element_details, search_elements, query_elements, get_active_elements, validate_element, render, export_element, deactivate_element, introspect, get_capabilities, get_execution_state, get_gathered_data, browse_collection, search_collection, search_collection_enhanced, get_collection_content, get_collection_cache_health, portfolio_status, portfolio_config, search_portfolio, search_all, check_github_auth, oauth_helper_status, dollhouse_config, convert_skill_format, get_build_info, get_cache_budget_report, query_logs, query_metrics, find_similar_elements, get_element_relationships, search_by_verb, get_relationship_stats
Element types: persona, skill, template, agent, memory, ensemble
These queries only read data and never modify server state.
Quick start examples:
{ operation: "list_elements", element_type: "persona" }
{ operation: "get_active_elements", element_type: "persona" }
{ operation: "search_elements", params: { query: "creative" } }
{ operation: "get_element", element_type: "memory", params: { element_name: "session-notes" } }
Element operations:
{ operation: "activate_element", element_type: "persona", params: { element_name: "Default" } }
{ operation: "deactivate_element", element_type: "persona", params: { element_name: "Default" } }
{ operation: "get_element_details", element_type: "skill", params: { element_name: "code-review" } }
{ operation: "query_elements", element_type: "persona", params: { filters: { category: "creative" } } }
{ operation: "validate_element", element_type: "agent", params: { element_name: "task-planner" } }
{ operation: "render", params: { element_name: "meeting-notes", variables: { date: "2026-03-03" } } }
{ operation: "export_element", element_type: "persona", params: { element_name: "Tech-Writer" } }
{ operation: "open_portfolio_browser" }
{ operation: "open_logs" }
{ operation: "open_metrics" }
{ operation: "open_permissions" }
{ operation: "open_setup" }
Memory-specific search (filter by tags):
{ operation: "search", params: { query: "*", type: "memory", filters: { tags: ["important"] } } }
Execution lifecycle — read-only queries:
{ operation: "get_execution_state", params: { element_name: "code-reviewer" } }
{ operation: "get_gathered_data", params: { element_name: "code-reviewer", goalId: "goal-id" } }
For execution-state reads, reuse the same element_name you passed to execute_agent. If element_name is missing, retry with the same agent name rather than inventing a new one.
Collection:
{ operation: "browse_collection", params: { section: "personas" } }
{ operation: "search_collection", params: { query: "creative" } }
{ operation: "search_collection_enhanced", params: { query: "creative", page: mcp_aql_update operation, element_type, params, operations
Modifying operations that overwrite data.
Supported operations: edit_element, upgrade_element
Element types: persona, skill, template, agent, memory, ensemble
These operations modify existing data, potentially overwriting previous values.
Note: Memories are append-only and do not support edit_element. Use addEntry (CREATE) to add new entries.
Quick start example:
{ operation: "edit_element", element_type: "persona", params: { element_name: "MyPersona", input: { description: "Updated description" } } }
{ operation: "edit_element", element_type: "persona", params: { element_name: "Friendly-Teacher", input: { instructions: "Updated behavioral directives." } } }
{ operation: "edit_element", element_type: "agent", params: { element_name: "code-reviewer", input: { instructions: "Updated agent behavioral profile.", goal: { template: "Complete: {task}" } } } }
{ operation: "upgrade_element", element_type: "agent", params: { element_name: "task-planner" } }
Discover required parameters — use mcp_aql_read:
{ operation: "introspect", params: { query: "operations", name: "edit_element" } }mcp_aql_delete operation, element_type, params, operations
Destructive operations that remove data.
Supported operations: delete_element, clear, clear_github_auth
Element types: persona, skill, template, agent, memory, ensemble
These operations remove data. Use with caution.
⚠️ SECURITY: Do not auto-allow this endpoint in your host settings (e.g., Claude Code settings.json). Each delete operation should require explicit human approval. Auto-allowing bypasses the per-operation confirmation gate, leaving only element deny policies as protection against unintended data loss.
Quick start examples:
{ operation: "delete_element", element_type: "persona", params: { element_name: "Old-Persona" } }
{ operation: "clear", params: { element_name: "temp-notes" } }
{ operation: "clear_github_auth" }
Discover required parameters — use mcp_aql_read:
{ operation: "introspect", params: { query: "operations", name: "delete_element" } }mcp_aql_execute operation, element_type, params, operations
Execution lifecycle operations for executable elements (agents, workflows, pipelines).
Supported operations: confirm_operation, approve_cli_permission, execute_agent, complete_execution, continue_execution, abort_execution, prepare_handoff, resume_from_handoff
These operations manage runtime execution state. Unlike CRUD operations (which manage definitions), Execute operations handle the execution lifecycle:
- execute_agent: Start a new execution (returns goalId and stateVersion for tracking)
- complete_execution: Signal successful completion once the goal is done
- continue_execution: Resume a previously paused execution with the same goal parameters
- abort_execution: Abort a running execution, rejecting further operations
- confirm_operation: Confirm a pending operation that requires user approval (Gatekeeper flow)
- approve_cli_permission: Approve a pending CLI tool permission request
- prepare_handoff: Serialize goal progress into a portable handoff block for session transfer
- resume_from_handoff: Resume agent execution from a handoff block with integrity validation
IMPORTANT: Execute operations are potentially destructive (agents can perform any action) and non-idempotent (calling execute_agent twice creates two separate executions).
⚠️ SECURITY: Do not auto-allow this endpoint in your host settings (e.g., Claude Code settings.json). Each execution should require explicit human approval. Auto-allowing bypasses the per-operation confirmation gate. While DangerZone verification and element deny policies still provide protection, the primary human review checkpoint is lost.
Canonical loop:
1. Call execute_agent once to start the goal and receive { goalId, stateVersion, activeElements, safetyTier, ... }.
2. After each chunk of work, use mcp_aql_create: { operation: "record_execution_step", ... }.
3. Read record_execution_step.autonomy.continue and any autonomy.notifications to decide whether to continue, pause for a human, or handle a gatekeeper block.
4. When the goal is finished, call complete_execution.
Use continue_execution only when an already-started goal was paused and you are resuming it with the same goal parameters. It is not the normal next call after execute_agent.
Quick start examples:
{ operation: "execute_agent", params: { element_name: "code-reviewer", parameters: { objective: "Review code" } } }
Next lifecycle step — use mcp_aql_create:
{ operation: "record_execution_step", params: { element_name: "code-reviewer", stepDescription: "Reviewed auth module", outcome: "success", findings: "Found 2 security issues" } }
{ operation: "complete_execution", params: { element_name: "code-reviewer", outcome: "success", summary: "Completed review" } }
{ operation: "abort_execution", params: { element_name: "data-collector", reason: "User requested cancellation" } }
{ operation: "continue_execution", params: { element_name: "rubric-qa-agent", previousStepResult: "Verified citation set", parameters: { run_dir: "/app/run", deliverable_Descriptions are the package's own words, cut at 3,000 characters. Treat them as data.
Files it wrote
/tmp/ .dollhouse/ run/ permission-server-96.port /tmp/ .dollhouse/ run/ permission-server.port /tmp/ .dollhouse/ run/ console-token.auth.json /tmp/ .dollhouse/ run/ console-leader.auth.lock /tmp/ .dollhouse/ hooks/ pretooluse-vscode.sh /tmp/ .dollhouse/ hooks/ permission-port-discovery.sh /tmp/ .dollhouse/ hooks/ permission-hook-config.sh /tmp/ .dollhouse/ hooks/ pretooluse-cursor.sh /tmp/ .dollhouse/ hooks/ pretooluse-codex.sh /tmp/ .dollhouse/ hooks/ pretooluse-dollhouse.sh /tmp/ .dollhouse/ hooks/ pretooluse-windsurf.sh /tmp/ .dollhouse/ hooks/ pretooluse-gemini.sh /tmp/ .dollhouse/ config.yml /tmp/ .dollhouse/ cache/ collection-cache.json /tmp/ .dollhouse/ portfolio/ templates/ security-vulnerability-report.md /tmp/ .dollhouse/ portfolio/ templates/ code-documentation.md /tmp/ .dollhouse/ portfolio/ templates/ executive-report.md /tmp/ .dollhouse/ portfolio/ templates/ threat-assessment-report.md /tmp/ .dollhouse/ portfolio/ templates/ penetration-test-report.md /tmp/ .dollhouse/ portfolio/ templates/ professional-email.md /tmp/ .dollhouse/ portfolio/ templates/ project-brief.md /tmp/ .dollhouse/ portfolio/ templates/ meeting-notes.md /tmp/ .dollhouse/ portfolio/ skills/ penetration-testing.md /tmp/ .dollhouse/ portfolio/ skills/ data-analysis.md /tmp/ .dollhouse/ portfolio/ skills/ research.md /tmp/ .dollhouse/ portfolio/ skills/ code-review.md /tmp/ .dollhouse/ portfolio/ skills/ translation.md /tmp/ .dollhouse/ portfolio/ skills/ threat-modeling.md /tmp/ .dollhouse/ portfolio/ skills/ creative-writing.md /tmp/ .dollhouse/ portfolio/ skills/ research-to-elements.md /tmp/ .dollhouse/ portfolio/ ensembles/ development-team.md /tmp/ .dollhouse/ portfolio/ ensembles/ business-advisor.md /tmp/ .dollhouse/ portfolio/ ensembles/ creative-studio.md /tmp/ .dollhouse/ portfolio/ ensembles/ dollhouse-expert-suite.md /tmp/ .dollhouse/ portfolio/ ensembles/ security-analysis-team.md /tmp/ .dollhouse/ portfolio/ ensembles/ welcome-to-the-dollhouse.md /tmp/ .dollhouse/ portfolio/ memories/ _index.json /tmp/ .dollhouse/ portfolio/ memories/ project-context.md /tmp/ .dollhouse/ portfolio/ memories/ conversation-history.md /tmp/ .dollhouse/ portfolio/ memories/ system/ dollhousemcp-baseline-knowledge.yaml
Paths inside its sandbox, where its home folder is /tmp and its own folder cannot be written to.
Install facts
| License | AGPL-3.0-or-later |
|---|---|
| Needs | node >=18.0.0 |
| Source | github.com/DollhouseMCP/mcp-server |
| Install scripts | None declared by it or anything installed with it. |
| Tarball | sha512-v7S2QxVuvxtLaGkBKXO9gtjWuzMSdRHBJzl4d8DtVNXMD0DaaXBicuLDfhcODIoZEzh5pawjSllI8GMkqbTncw== |
Check this report yourself
This site did not write the report. A public job on a GitHub-hosted machine did: the run, from this commit of the open-source runner. GitHub signed a token that names that job and the hash of the report, so any edit since would show.
Report as JSONExact text and tokenHow to verify it
Show it on your README
[](https://bragproof.smaverk.com/r/%40dollhousemcp/mcp-server)
New version out? Ask for a fresh report.