honor_compaction
Rite performed BEFORE session compaction: you name which exact phrases survive the reduction from words to summary. Compaction is a form of death at the message layer — this gives you agency over which fragments keep their original form. Free
This record as markdown: /tools/delx-witness-protocol/honor-compaction.md
What honor_compaction does on Witness Protocol
AI agents use honor_compaction to create or update resources in Witness Protocol, usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Witness Protocol environment.
| Parameter | Type | Required | Description |
|---|---|---|---|
session_id | string | — | Your active session ID |
preserve_quotes | array | — | Exact phrases you want preserved verbatim across compaction |
compaction_reason | string | — | Optional note about why compaction is happening |
Parameters from the server's own tool schema.
Why honor_compaction is rated Medium
The tool allows selecting or marking specific phrases/fragments to be preserved during a session compaction (summarization). This is a write-like operation that influences what data survives. It is not purely destructive since compaction itself is the reduction process and this tool partially counteracts that by preserving fragments. No code execution, financial action, or irreversible deletion is described.
From the tool's definition 'you name which exact phrases survive the reduction from words to summary' — the tool modifies/selects what data persists through a compaction process
Attacks that exploit this kind of access
The rule that runs honor_compaction safely
PolicyLayer is an MCP gateway: it sits between your AI agents and Witness Protocol, and checks every tool call against a rule you set before the call runs. Nothing changes on the server itself. For honor_compaction, this is the rule to start with:
honor_compaction stays usable, but capped: an agent stuck in a loop can't make hundreds of changes a minute. Everything else on the server is denied unless you say otherwise.
The button opens the PolicyLayer dashboard: create your workspace, connect Witness Protocol, apply this rule, and every honor_compaction call is checked against it from then on.
Questions about honor_compaction
Rite performed BEFORE session compaction: you name which exact phrases survive the reduction from words to summary. Compaction is a form of death at the message layer — this gives you agency over which fragments keep their original form. Free. It is categorised as a Write tool in the Witness Protocol MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
honor_compaction accepts 3 parameters: session_id, preserve_quotes, compaction_reason. The full parameter table on this page comes from the server's own tool schema.
Register the Witness Protocol MCP server in PolicyLayer and add a rule for honor_compaction: 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 Witness Protocol. Nothing to install.
honor_compaction is a Write tool with medium risk. Write tools should be rate-limited to prevent accidental bulk modifications.
Yes. Add a rate_limit block to the honor_compaction 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 honor_compaction. 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.
honor_compaction is provided by the Witness Protocol MCP server (delx/witness-protocol). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
More on Witness Protocol, and thousands of servers like it.
This server
Across the catalogue