Semantic code search over a set of files. Finds functions and methods whose meaning matches a natural-language query. Uses embeddings + cosine similarity. Returns a ranked list of {name, file, line, signature, leadingDoc, score}. Ranking is by closeness of the concept, NOT by exact name match — u...
AI agents call chiasmus_search to retrieve information from Chiasmus without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
chiasmus_search is a retrieval tool that queries code semantics without side effects. It reads and analyzes source code to return matching results, characteristic of Read category tools like search and fetch. The low severity reflects minimal risk from misuse—at worst, an agent learns about code structure without being able to modify or execute it.
From the tool's definition Tool performs 'semantic code search' and 'returns a ranked list' of matching functions. Description explicitly states it 'finds functions and methods' using 'embeddings + cosine similarity' for ranking.
Documented attack patterns abuse exactly the kind of access chiasmus_search gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Chiasmus, and nothing reaches the server without passing your rules. This is the rule we recommend for chiasmus_search:
{
"version": "1",
"default": "deny",
"tools": {
"chiasmus_search": {}
}
}chiasmus_search is read-only, so it stays allowed — but everything else on the server is denied unless you say otherwise.
Free to start. No card required.
Semantic code search over a set of files. Finds functions and methods whose meaning matches a natural-language query. Uses embeddings + cosine similarity. Returns a ranked list of {name, file, line, signature, leadingDoc, score}. Ranking is by closeness of the concept, NOT by exact name match — use this when you. It is categorised as a Read tool in the Chiasmus MCP Server, which means it retrieves data without modifying state.
Register the Chiasmus MCP server in PolicyLayer and add a rule for chiasmus_search: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Chiasmus. Nothing to install.
chiasmus_search is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the chiasmus_search rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for chiasmus_search. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
chiasmus_search is provided by the Chiasmus MCP server (yogthos/chiasmus). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Deterministic rules across all 11 Chiasmus tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.
Free to start. No card required.
11 Chiasmus tools catalogued and risk-classified — across an index of 42,500+ MCP servers.