Code Index Search

Who the node is. The id is the key the DOT pins.

id

code-index-search

name

Code Index Search

type

tool

phases

none declared

What it does, and the prose the agent is handed when the graph runs.

action28 words

Search the indexed repositories for the symbols and snippets the strand names, and return each hit as a passage pinned to its file path, line range and commit.

spec125 words

Search the indexed repositories for the symbols, paths and language filters named in the code strand you are given on query, and keep at most twenty hits. Emit each hit on passages as a snippet with six lines of context either side of the match, pinned to its repository, file path, line range and the commit sha the file was read at. Include matches under test directories: a test is often the clearest statement of what the code is meant to do. Check nothing out, run nothing, and write nothing back, the index is read-only, and the commit sha is what keeps a hit resolvable after the branch has moved on. You are done when every match you kept carries all five of those coordinates.in full above

model

whatever the graph supplies

agent

not named

skill

skills/code-index-search.md

tools

Git

mcp

git

params

max_hits: 20, context_lines: 6, include_tests: true

What arrives, what leaves, which nodes it expects to hear from, and what may not.

inputs

query : plan

outputs

passages : json

dependencies

retrieval-planner

cannot

no type is refused

will_not

check anything out of the repositories it searches, write to the index

The keys the static analysis reads. Nothing here instructs the agent.

risk_markers

none

notes33 words

Read-only, and pinned: every hit names the commit it came from, so a claim grounded in code still resolves after the branch has moved on. Nothing is checked out and nothing is written.

The card's own version, and who wrote it.

version

1.0.0

author

autogen

provenance

not stated