Built-in Tools

Open Astra ships with 67 built-in tools. All are available to any agent subject to the agent's allow/deny policy.

File System

Tool	Description
`file_read`	Read the contents of a file at a given path
`file_write`	Write content to a file, creating it if it does not exist
`file_append`	Append content to an existing file
`file_delete`	Delete a file or directory
`file_list`	List files in a directory with optional glob filtering
`file_search`	Search for files matching a pattern recursively

Shell and Process

Tool	Description
`shell_exec`	Execute a shell command and return stdout/stderr
`shell_exec_long`	Execute a long-running command with streaming output
`process_list`	List running processes

Web and HTTP

Tool	Description
`web_search`	Search the web via Brave Search API and return summarized results
`web_fetch`	Fetch the content of a URL and return its text
`http_request`	Make an arbitrary HTTP request (GET, POST, PUT, DELETE)
`graphql-query`	Execute a GraphQL query or mutation with optional variables, custom headers, and schema introspection

Code

Tool	Description
`code_search`	Search the codebase using ripgrep-style patterns
`codebase_index`	Index the codebase for semantic search queries
`codebase_search`	Semantic search over indexed codebase chunks
`git_command`	Run any git command and return its output
`git_diff`	Get the diff for staged, unstaged, or a specific commit

Memory

Tool	Description
`memory_write`	Write a fact or note to the user's personal memory (Tier 2)
`memory_search`	Search across all memory tiers for a query
`workspace_memory_write`	Write a team-scoped entry to workspace memory
`workspace_memory_search`	Search workspace memory entries
`workspace_knowledge_write`	Write an entry to the structured knowledge base
`workspace_knowledge_search`	Search the knowledge base
`graph_write`	Write an entity or relationship to the knowledge graph
`graph_traverse`	Traverse the knowledge graph from a seed entity
`document_search`	Search uploaded RAG documents by semantic similarity
`memory_compress`	On-demand tier-specific memory compression with diff output
`memory_propagate`	Hierarchical cross-tier promotion of memory entries
`memory_budget`	Per-agent memory access limits with enforcement
`consistency_check`	Cross-tier contradiction detection across memory layers
`memory_feedback`	Post-synthesis utility feedback for memory retrieval tuning
`relation_decay`	Manual time-based graph edge strength decay management

Agents and Orchestration

Tool	Description
`spawn_agent`	Spawn a sub-agent with a task and shared blackboard context
`blackboard_write`	Write a value to the swarm blackboard
`blackboard_read`	Read a value from the swarm blackboard
`blackboard_list`	List all keys on the current swarm blackboard
`blackboard_topic`	Named pub/sub channels for selective swarm communication
`code_review_swarm`	Multi-agent autonomous code review with specialized reviewers
`swarm_observe`	Read-only swarm monitoring with alerts
`meta_controller`	Multi-swarm overseer with resource allocation
`debate_witness`	Evidence retrieval from memory for conflict resolution

Session and Scenario

Tool	Description
`session_branch`	Fork sessions into parallel experiment branches
`scenario_engine`	What-if simulation with side-by-side comparison

Multi-Modal and Routing

Tool	Description
`multimodal_input`	Unified multi-modal input processing (text, image, audio)
`provider_route`	Cross-provider intelligent routing based on task requirements
`model_inference`	Direct model inference call with full parameter control (temperature, top-p, stop sequences)

Deep Research and Dream Mode

Tool	Description
`research_manage`	Start, iterate, and retrieve results from long-running deep research tasks
`dream_manage`	Trigger and retrieve results from dream mode offline simulation sessions

Database

Tool	Description
`database_query`	Execute parameterized queries against the connected Postgres database
`graph_schema_evolution`	Add, modify, or deprecate knowledge graph entity and edge types

Agent Self-Improvement

Tool	Description
`ethical_check`	Post-generation ethical alignment validation with category scoring
`agent_therapy`	Self-reflection journal and behavioral pattern analysis
`health_scorecard`	Periodic agent reliability and efficiency scoring
`evolution_feedback`	Goal-driven persona evolution with feedback loops

Workspace

Tool	Description
`prompt_template`	Versioned reusable prompt templates
`knowledge_inject`	External knowledge injection with deduplication

Utilities

Tool	Description
`result_summarizer`	Summarize a long piece of text to fit within a token budget
`calculator`	Evaluate mathematical expressions safely
`date_time`	Get current date and time in any timezone
`json_transform`	Apply jq-style transformations to JSON data
`text_diff`	Compute a readable diff between two strings
`pdf_read`	Extract text content from a PDF file or URL