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

With no annotations present, the description carries the full burden and does meaningful work: it discloses the token budget of the default call (~500-800 tokens), the lazy hierarchical retrieval model (matches project → branch → session first, drills into raw chunks only when needed), and the depth control. It does not state permission/auth requirements or explicitly confirm the operation is read-only, though 'Retrieve' strongly implies it.

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?

Well front-loaded: the core action and the recommended no-argument session-start call come first, followed by query mode and depth. It is efficient, but the session-start instruction is stated twice ('Call at session start with no arguments' and 'Always call this first in a new session'), which is mild redundancy.

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?

For a 4-parameter tool with no annotations and no output schema, the description is nearly sufficient: it explains the default invocation, the search mode, depth progression, and roughly what gets returned and how large it is. The remaining gaps are the response shape of query results and any auth/prerequisite context, neither of which is covered anywhere else.

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?

Schema description coverage is 100%, so the schema already documents depth levels, the query hierarchy, the type filter, and the max_tokens default — baseline 3 applies. The description echoes the depth and query behavior but adds no syntax, format, or interaction detail beyond what the schema fields state.

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?

States a specific verb and resource ('Retrieve persistent memory from previous sessions') and describes the concrete artifact returned: the context pyramid of project overview, branch summary, and recent session summaries. It implicitly separates itself from the write-oriented sibling save_context by framing itself as the read/start-of-session tool, but it never names get_status or summarize_session, so sibling differentiation is only partial.

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?

Gives clear conditions for the two main modes — 'Call at session start with no arguments' for the pyramid, and 'Use with a query' for hierarchical search — plus 'Use depth to control how deep to search.' It even gives ordering guidance ('Always call this first in a new session'). It stops short of naming when NOT to use it or pointing at the alternatives (get_status, summarize_session), which is what a 5 would require.

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

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

No annotations are provided, so the description carries the burden. It implicitly reveals a safe, read-only, zero-argument inspection operation and discloses the exact fields returned, which is useful given there is no output schema. It does not mention permission requirements, cost, or freshness guarantees for the reported values, so a moderate score is warranted.

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?

A single front-loaded sentence that gets to the point immediately, and every listed item is informative because there is no output schema to enumerate them. Slightly list-heavy but none of it is wasted.

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?

For a zero-parameter, no-output-schema status tool, the description adequately covers what the agent gets back (the enumerated stats) and how to treat the call. Missing only when to prefer it over the sibling context tools.

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?

The tool takes zero parameters, so there is nothing for the description to disambiguate; baseline is 4. This is the correct ceiling for a no-argument tool.

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 states a specific resource (cortexmem) and enumerates exactly what the tool reports: total chunks, type breakdown, storage location, last init time, last indexed commit. It's clearly a read-only status/inspection tool, distinguishable from the save/get/summarize context siblings. No sibling is named explicitly, so it stops short of a 5.

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?

There is no statement of when to call this tool versus the alternatives (save_context, get_context, summarize_session), nor any prerequisites or exclusions. Usage is only implied by the fact that it reports status.

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

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

With no annotations, the description must carry the full behavioral burden. It usefully discloses that saved context 'persists across sessions and editors' and that calling proactively matters. However, it says nothing about overwrite/deduplication behavior, whether saves are keyed by type, permissions, or what a save returns — notable gaps for a mutation tool with no annotations.

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?

Three tight sentences, front-loaded with the core action, then usage triggers, then the persistence rationale. No filler; every clause earns its place.

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?

For a four-parameter write tool with no output schema and no annotations, the description covers what it does, when to call it, and the cross-session persistence guarantee. It is nearly complete, missing only the post-save mechanics (what happens to existing context, what is returned).

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?

Schema description coverage is 100%, so the schema already documents all four parameters including the enum mappings and an inline example. The description's prose enumeration of the context types mirrors the enum rather than extending it, so baseline 3 applies.

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?

States a specific verb and resource: 'Save context to persistent memory.' The write-side counterpart to siblings like get_context is obvious from the verb, though the description never names a sibling to sharpen the boundary. Purpose is unambiguous and distinct from the read/summary siblings.

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?

Gives concrete, enumerated triggers — decisions, non-obvious discoveries, agreed constraints, WIP state, coding preferences — plus an explicit 'call proactively' directive. It lacks any 'when not to use' guidance or a pointer to the retrieval sibling (get_context), so it stops short of full routing guidance.

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

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

With no annotations, the description carries the full burden and does well: it discloses that three persisted artifacts are created/updated, that an ANTHROPIC_API_KEY enables LLM compaction, and that a deterministic fallback exists without it. It omits whether repeated calls are idempotent or how the writes can be undone.

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?

Four short sentences, front-loaded with the action and its outputs, followed by the trigger condition and the auth/fallback caveat. Every sentence contributes distinct information with no padding.

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?

For a mutation tool with no annotations and no output schema, the description covers side effects, timing, and the auth-dependent execution path well. It does not describe what the agent gets back on success or failure, which is the remaining gap.

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?

Only one optional parameter at 100% schema description coverage, so the schema already defines it fully. The description adds no format, length, or content guidance for session_summary beyond noting context comes from 'saved context,' so the baseline 3 applies.

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?

States a specific verb+resource ('Compact and persist session memory') and then enumerates the three concrete artifacts it produces (session summary, branch summary, project overview), which clearly distinguishes it from save_context/get_context/get_status.

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?

Explicitly says 'Call at end of session,' giving a clear usage trigger. It does not name a when-not condition or point at an alternative sibling for related memory operations, so it stops short of full routing guidance.

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