Skip to main content

Module tools

Module tools 

Source

Modules§

a2a
activate_skill
apply_patch
apply_patch: add, delete, move and edit several files in one call.
background_process
Background shell processes for long-running commands.
bash
browser
clarify
consult_specialist
deep_research
deep_research — a deterministic multi-source research gatherer.
describe_image
edit_file
explore
explore: a synchronous, read-only sub-agent for wide searches.
export_chat
fetch_artifact
file_state
Stale-read guard shared by read_file, edit_file and write_file.
fuzzy_match
Multi-strategy fuzzy matching for edit_file.
generate_image
generate_podcast
glob
grep
insights
knowledge_graph
learning_tracks
mcp
memory
osv_check
read_file
render_pdf
report_progress
schedule
send_message
session_search
skill_manage
specialists
Specialist sub-agent profiles.
structured_memory
subagents
sync_skills
text_to_speech
time_math
todo
transcribe_audio
web_fetch
web_search
workspace_walk
Gitignore-aware directory traversal shared by the grep and glob tools.
write_file

Structs§

ToolAuthContext
ToolRegistry
ToolResult

Enums§

ToolRisk
Declaration order is the risk ordering (Low < Medium < High) — the derived Ord is relied on by the tool-policy max_risk comparison.

Constants§

IDEMPOTENT_TOOLS
Tools that are read-only / side-effect-free for the same arguments. Used by the per-turn guardrail controller (tool_guardrails.rs) to detect “no progress” loops where the model keeps re-running the same query and getting the same result. NOT a security boundary — that’s tool_risk / tool_execution_policy in the runtime crate.

Traits§

Tool

Functions§

auth_context_from_input
authorize_chat_access
resolve_tool_path
resolve_tool_working_dir
schema_object
tool_execution_policy
tool_risk
validate_execution_policy