mcp_engram_forget_old

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations provided, the description carries the full burden of behavioral disclosure. It thoroughly details what gets evicted (non-pinned, below threshold), what is exempt (pinned blocks at CRS=1.0), the Langevin ranking formula, batching via max_evict, and a safety recommendation to preview with stats first. This goes well beyond a minimal description.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is moderately long but each sentence contributes valuable information, including usage guidance, exceptions, algorithm details, and a caution to preview first. It is well-structured with a clear 'WHEN TO USE' section, though it could be slightly tightened without losing critical content.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

The description covers the tool's purpose, usage, parameters, and core behavior, which is substantial especially since there is no output schema. However, it does not explicitly mention what the tool returns (e.g., count of evicted blocks) or whether eviction is reversible, which would be valuable for a destructive operation.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Although the schema already describes each parameter (100% coverage), the description adds substantial semantic value: it explains the default threshold (0.2) and suggests starting at 0.3, clarifies that older_than_days targets stale memories while preserving recent ones, and details the Langevin ranking formula for langevin_rank. This significantly enhances understanding beyond the schema.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states that the tool evicts non-pinned memories below a CRS threshold, with a specific verb ('evict'), resource ('non-pinned memories'), and scope ('below CRS threshold'). It explicitly distinguishes itself from automatic garbage collection and from mcp_engram_stats (for preview), providing a unique purpose.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description includes a dedicated 'WHEN TO USE' section with concrete scenarios (end of project phase, after distill runs, manifold growing too large) and offers practical advice (start with threshold 0.3, preview with mcp_engram_stats). However, it does not explicitly state when NOT to use this tool or compare directly to the sibling mcp_engram_forget.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.