Tools Reference

Every id below is a function id valid in the tools.list field of an agent YAML (the standalone agent runtime that powers Slack Bot, Programmable, and Bolt Pipeline agents). Use allowlist mode to restrict an agent to only what it needs, or denylist to remove tools from the default set. See YAML Configuration for syntax.

No tools block = full tool access. Side-effectful tools (for example trigger_bolt_schedule_run and the Bolt retries) are never available unless named explicitly in an allowlist. Overriding a schedule’s commands or branch is a separate tool, trigger_bolt_schedule_run_with_overrides: an agent that allowlists only trigger_bolt_schedule_run and passes commands to it is refused. Some tools are Copilot-only (DuckDB, GitHub PR) and are not available to agents, so they do not appear here.

Always-available tools

Nine core tools are available to every agent regardless of its tools block: an allowlist does not need to name them, and a denylist cannot remove them. This means an allowlist can no longer strand an agent in plan mode or silently disable Agent Skills.

Id	Description
`read_file`	Read a file in the repo
`search_files_and_directories`	List and navigate the repo file structure
`ripgrep_search`	Search code patterns across files with ripgrep
`enter_plan_mode` / `exit_plan_mode`	Enter and leave plan mode
`load_skill_instructions`	Load instructions for a named skill
`run_subagent`	Run a lightweight sub-task without a full agent pod
`search_paradime_docs` / `get_paradime_doc_page`	Search and read the Paradime help docs

Sub-agents started with run_subagent inherit the parent agent’s tool gating, so a tightly allowlisted agent cannot reach extra tools by delegating. The explicit-allowlist-only tools (the Bolt trigger and retry tools, and the Slack thread tools) are never available to sub-agents, even when the parent allowlists them.

Filesystem & code

File System Tool

Id	Description
`read_file`	Read a file in the repo, up to 2,000 line-numbered lines per call, with `offset` and `limit` to page through a large file
`edit_file`	Replace an exact string in a file, so an edit costs the size of the change rather than the size of the file
`write_file`	Create or overwrite a file
`search_files_and_directories`	List and navigate the repo file structure

edit_file comes with write_file. An agent that allowlists write_file is offered edit_file too, and denying write_file removes both, so existing agents get surgical edits without an allowlist change. An edit is refused unless the agent has already read the current version of the file in that run.

Code Search Tool

Id	Description
`ripgrep_search`	Search code patterns across files with ripgrep

Terminal Tool

Id	Description
`run_terminal_command`	Run one command without a shell (dbt™, git, gh, glab, or az)

Agent-only (no dedicated tool page)

Id	Description
`load_skill_instructions`	Load instructions for a named skill

Web, search & docs

Web Search Tool

Id	Description
`execute_web_search`	Search the web
`extract_url_content`	Fetch and read a URL

Perplexity Tool

Id	Description
`perplexity_search`	AI-powered web search via Perplexity

Paradime Docs Tool

Id	Description
`search_paradime_docs`	Search the Paradime help docs
`get_paradime_doc_page`	Read a Paradime help docs page

Warehouse: SQL & metadata

SQL Execution Tool

Id	Description
`run_sql_query`	Execute a SQL query against the connected warehouse
`get_sql_query_result`	Collect the result of a Snowflake query that `run_sql_query` left running (Snowflake only)
`cancel_sql_query`	Stop a Snowflake query that `run_sql_query` started (Snowflake only)

On Snowflake, an agent’s run_sql_query waits up to 60 seconds for the result. A query that is still running after that keeps running in Snowflake, and the tool returns its query_id. The agent can do other work, then collect the rows with get_sql_query_result or stop the query with cancel_sql_query. Snowflake stops the query on its own after 15 minutes, and Paradime cancels the session’s running queries when the session stops. An agent can only collect or cancel queries that its own session started. An allowlist that names run_sql_query gets both tools without naming them. On other warehouses, a query must finish within the 5-minute tool limit. Snowflake Tools

Id	Description
`list_all_snowflake_databases`	List databases
`list_all_schemas_in_snowflake_database`	List schemas
`list_all_tables_in_snowflake_schema`	List tables
`list_all_columns_in_snowflake_table`	List columns
`get_snowflake_query_performance_stats`	Analyze a query’s performance

BigQuery Tools

Id	Description
`list_all_bigquery_projects_and_datasets`	List projects and datasets
`list_all_tables_in_bigquery_dataset`	List tables
`list_all_columns_in_bigquery_table`	List columns
`get_bigquery_query_performance_stats`	Analyze a job’s performance

Redshift Tools

Id	Description
`list_all_redshift_databases`	List databases
`list_all_schemas_in_redshift_database`	List schemas
`list_all_tables_in_redshift_schema`	List tables
`list_all_columns_in_redshift_table`	List columns

Trino Tools

Id	Description
`list_all_trino_databases`	List databases
`list_all_schemas_in_trino_database`	List schemas
`list_all_tables_in_trino_schema`	List tables
`list_all_columns_in_trino_table`	List columns

Databricks Tools

Id	Description
`list_all_databricks_databases`	List databases
`list_all_schemas_in_databricks_database`	List schemas
`list_all_tables_in_databricks_schema`	List tables
`list_all_columns_in_databricks_table`	List columns

SQL Server Tools

Id	Description
`list_all_sqlserver_databases`	List databases
`list_all_schemas_in_sqlserver_database`	List schemas
`list_all_tables_in_sqlserver_schema`	List tables
`list_all_columns_in_sqlserver_table`	List columns

ClickHouse Tools

Id	Description
`list_all_clickhouse_databases`	List databases
`list_all_tables_in_clickhouse_database`	List tables
`list_all_columns_in_clickhouse_table`	List columns

Microsoft Fabric Tools

Id	Description
`list_all_fabric_databases`	List databases
`list_all_schemas_in_fabric_database`	List schemas
`list_all_tables_in_fabric_schema`	List tables
`list_all_columns_in_fabric_table`	List columns

MongoDB Tools

Id	Description
`list_mongodb_databases`	List databases
`list_mongodb_collections`	List collections
`mongodb_collection_schema`	Infer a collection’s schema
`mongodb_find`	Run a read-only find query
`mongodb_count`	Count documents
`mongodb_aggregate`	Run a read-only aggregation pipeline
`mongodb_collection_indexes`	Describe collection indexes
`mongodb_db_stats`	Return database storage stats

dbt Discovery

dbt Discovery Tools Read-only project metadata from your Bolt runs’ dbt™ artifacts. Names and output mirror the dbt-mcp Discovery toolset.

Id	Description
`get_all_models`	List all models with descriptions
`get_mart_models`	List mart models only
`get_all_sources`	List sources with freshness status
`get_all_macros`	List macros (builtin packages excluded by default)
`get_exposures`	List downstream dashboards, apps, and analyses
`get_lineage`	Lineage graph around a node at model grain: which models depend on this model. For a single column’s value path, use `get_column_level_lineage` below
`get_model_health`	Last run status, test statuses, and upstream health for a model
`get_model_performance`	Historical run timings for a model
`get_node_details`	Full details for any model, source, exposure, test, seed, snapshot, or macro

Catalog & lineage

Catalog Tool

Id	Description
`search_catalog`	Search the data catalog (dbt™ assets and connected integrations) by name, description, tag, or column. Takes `query` and `limit` (default 5, max 25). A dbt™ result’s `key` is its `unique_id`

Column Level Lineage

Id	Description
`get_column_level_lineage`	Trace one column’s value path at column grain: where it comes from, and what breaks if you change it. Takes `upstream_depth` and `downstream_depth`

These read Paradime’s own catalog and lineage graph rather than the dbt™ Discovery API, so they are their own group. Scope is the production environment plus connected integrations, including dbt™ Mesh.

Bolt

Bolt Tool

Id	Description
`list_bolt_schedules`	List active Bolt schedules
`get_bolt_run_logs`	Fetch logs and summaries from a run
`trigger_bolt_schedule_run`	Run a schedule exactly as configured (explicit allowlist only)
`trigger_bolt_schedule_run_with_overrides`	Run a schedule with overridden commands or branch, for that run only (explicit allowlist only)
`retry_bolt_run`	Retry a specific Bolt run (explicit allowlist only)
`retry_latest_bolt_schedule_run`	Retry the latest run of a schedule (explicit allowlist only)

Integrations

Jira Tool

Id	Description
`get_jira_issue`	Fetch an issue
`list_jira_issues`	List issues
`create_jira_issue`	Create an issue
`add_jira_watchers`	Add watchers to an existing issue
`link_jira_issues`	Link two existing issues (blocks, is blocked by, relates to, duplicates, or any link type in your site)
`update_jira_issue`	Update or transition an issue
`add_jira_comment`	Add a comment
`export_jira_issues_to_csv`	Export issues to CSV
`attach_slack_file_to_jira_issue`	Attach a file shared in the session’s Slack conversation to an issue

HubSpot Tool

Id	Description
`search_hubspot_objects`	Search records of any object type
`get_hubspot_object`	Fetch a single record
`list_hubspot_properties`	List the properties defined for an object type
`list_hubspot_owners`	List HubSpot owners
`get_hubspot_associations`	Read the records associated with a record
`create_hubspot_object`	Create a record
`update_hubspot_object`	Update a record
`add_hubspot_note`	Add a note and attach it to a record

Linear Tool

Id	Description
`get_linear_issue`	Fetch an issue
`list_linear_issues`	List issues
`create_linear_issue`	Create an issue
`update_linear_issue`	Update an issue
`add_linear_comment`	Add a comment

Confluence Tool

Id	Description
`get_confluence_page`	Read a page
`create_confluence_page`	Create a page
`update_confluence_page`	Update a page
`create_confluence_footer_comment`	Add a footer comment
`create_confluence_inline_comment`	Add an inline comment

Notion Tool

Id	Description
`search_notion`	Search pages and databases
`fetch_notion_page`	Fetch a page
`get_notion_page_comments`	Get page comments

Google Drive Tool

Id	Description
`search_google_drive_files`	Search Google Drive

Google Docs Tool

Id	Description
`read_google_doc`	Read a Google Doc

Google Sheets Tool

Id	Description
`read_google_sheet`	Read a Google Sheet

Slack Tools

Id	Description
`find_slack_thread`	Search a channel for an existing thread to attach to (explicit allowlist only). Limited to the agent’s own channels, see below
`attach_session_to_slack_thread`	Attach the session to a Slack thread (explicit allowlist only). Limited to the agent’s own channels, see below
`post_slack_message`	Post a message to the session’s Slack thread or channel
`upload_file_to_slack`	Upload a file to Slack

find_slack_thread and attach_session_to_slack_thread work only in the agent’s own channels: the slack.channel in its definition, and the channel its session is bound to. An agent with neither can use them in any channel the Paradime bot is in.

Orchestration & repositories

Agent-only (no dedicated tool page)

Id	Description
`run_subagent`	Run a lightweight sub-task without a full agent pod
`invoke_agent`	Spawn a child agent defined in agents_squad
`notify_parent_session`	Wake a parent agent waiting on a callback
`check_delegated_agents`	Report the status and latest output of the agents this session delegated to
`list_accessible_repos`	List repositories the agent can access
`clone_additional_repo`	Clone another repository into the workspace

Task list

Agent-only (no dedicated tool page)

Id	Description
`todo_write`	Write or update the run’s checklist
`todo_read`	Read the run’s checklist back

A multi-step run keeps its plan on the session rather than re-deriving it in context every turn.

Plan mode

Agent-only (no dedicated tool page)

Id	Description
`enter_plan_mode`	Enter plan mode (read-only planning)
`exit_plan_mode`	Exit plan mode and begin executing

Code Review

Slack Bot Agent

Programmable Agent

Bolt Pipeline Agent

Copilot

MCP Server

Context & Customization

DinoAI Tools

Settings & Credits

Always-available tools

Filesystem & code

Web, search & docs

Warehouse: SQL & metadata

dbt Discovery

Catalog & lineage

Bolt

Integrations

Orchestration & repositories

Task list

Plan mode

​Always-available tools

​Filesystem & code

​Web, search & docs

​Warehouse: SQL & metadata

​dbt Discovery

​Catalog & lineage

​Bolt

​Integrations

​Orchestration & repositories

​Task list

​Plan mode

Always-available tools

Filesystem & code

Web, search & docs

Warehouse: SQL & metadata

dbt Discovery

Catalog & lineage

Bolt

Integrations

Orchestration & repositories

Task list

Plan mode