Find context relevant to a specific user request. Use this when you need context that's specifically relevant to the current task, rather than all recent context. This is more token-efficient for large projects. **IMPORTANT: Call this AFTER initial exploration of the request**, not immediately....
Single-target operation
Part of the Kawa Code MCP MCP server. Enforce policies on this tool with Intercept, the open-source MCP proxy.
AI agents call get_relevant_context to retrieve information from Kawa Code MCP without modifying any data. This is common in research, monitoring, and reporting workflows where the agent needs context before taking action. Because read operations don't change state, they are generally safe to allow without restrictions -- but you may still want rate limits to control API costs.
Even though get_relevant_context only reads data, uncontrolled read access can leak sensitive information or rack up API costs. An agent caught in a retry loop could make thousands of calls per minute. A rate limit gives you a safety net without blocking legitimate use.
Read-only tools are safe to allow by default. No rate limit needed unless you want to control costs.
tools:
get_relevant_context:
rules:
- action: allowSee the full Kawa Code MCP policy for all 17 tools.
Agents calling read-class tools like get_relevant_context have been implicated in these attack patterns. Read the full case and prevention policy for each:
Other tools in the Read risk category across the catalogue. The same policy patterns (rate-limit, allow) apply to each.
Find context relevant to a specific user request. Use this when you need context that's specifically relevant to the current task, rather than all recent context. This is more token-efficient for large projects. **IMPORTANT: Call this AFTER initial exploration of the request**, not immediately. Once you know which files are involved, pass them in `activeFiles` for much better relevance matching. A vague prompt alone yields poor results. The tool: 1. Sends your prompt to the API for semantic embedding 2. Computes cosine similarity against stored embeddings for all intents and decisions 3. Uses `activeFiles` to boost items with matching file paths 4. Returns only the most relevant items above the minimum score threshold **Recommended workflow:** 1. `check_active_intent` at SESSION START (resume existing work) 2. Explore the user's request (read files, understand scope) 3. `get_relevant_context` with prompt + activeFiles (task-specific context) Returns: - **relevantIntents**: Past work related to the current task - **relevantDecisions**: Decisions affecting similar code/concepts (both intent-scoped and repo-scoped) Example: User asks "Add validation to user registration endpoint" After exploring, you found src/routes/user.ts and src/validators/ Call with: prompt + activeFiles: ["src/routes/user.ts", "src/validators/user.ts"] Returns intents/decisions matching these files and keywords. It is categorised as a Read tool in the Kawa Code MCP MCP Server, which means it retrieves data without modifying state.
Add a rule in your Intercept YAML policy under the tools section for get_relevant_context. You can allow, deny, rate-limit, or validate arguments. Then run Intercept as a proxy in front of the Kawa Code MCP MCP server.
get_relevant_context 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 get_relevant_context rule in your Intercept 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 Intercept policy for get_relevant_context. 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.
get_relevant_context is provided by the Kawa Code MCP MCP server (@kawacode/mcp). Intercept sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Open source. One binary. Zero dependencies.
npx -y @policylayer/intercept