pack_context
Pack project context into a single document for external LLMs. Intelligent selection by graph importance, fits within token budget. Better than Repomix for focused context. Strategies: most_relevant (default — feature/PageRank ranked), core_first (PageRank always wins, surfaces architecturally ce...
This record as markdown: /tools/nikolai-vysotskyi-trace-mcp/pack-context.md
What pack_context does on Trace
AI agents call pack_context to retrieve information from Trace without modifying anything. It is typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
Why pack_context is rated Low
This tool retrieves and aggregates source code information for export purposes. It performs no mutations, executions, or destructive operations. The explicit 'Read-only' designation and the absence of any side effects confirm it is a data retrieval operation. While context packing could theoretically expose sensitive code, the act itself is non-invasive information retrieval.
From the tool's definition Pack project context into a single document for external LLMs. Intelligent selection by graph importance, fits within token budget. Read-only.
Attacks that exploit this kind of access
The rule that runs pack_context safely
PolicyLayer is an MCP gateway: it sits between your AI agents and Trace, and checks every tool call against a rule you set before the call runs. Nothing changes on the server itself. For pack_context, this is the rule to start with:
pack_context is read-only, so it stays allowed. Everything else on the server is denied unless you say otherwise.
The button opens the PolicyLayer dashboard: create your workspace, connect Trace, apply this rule, and every pack_context call is checked against it from then on.
Questions about pack_context
Pack project context into a single document for external LLMs. Intelligent selection by graph importance, fits within token budget. Better than Repomix for focused context. Strategies: most_relevant (default — feature/PageRank ranked), core_first (PageRank always wins, surfaces architecturally central code), compact (signatures only — drops source bodies, lets outlines cover much more of the repo per token). Read-only. Use when sharing project context with external tools. Returns XML/Markdown/JSON with selected code within budget. It is categorised as a Read tool in the Trace MCP Server, which means it retrieves data without modifying state.
Register the Trace MCP server in PolicyLayer and add a rule for pack_context: 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 Trace. Nothing to install.
pack_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 pack_context 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 pack_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.
pack_context is provided by the Trace MCP server (nikolai-vysotskyi/trace-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
More on Trace, and thousands of servers like it.
This server
Across the catalogue