Recall relevant memory artifacts (lessons, patterns, reasoning traces) for the current task. Hybrid semantic + full-text ranking — finds conceptually related memory, not just keyword matches.
| Name | Required | Description | Default |
|---|---|---|---|
| limit | No | ||
| query | Yes | Describe the current task or problem | |
| workspace | No | Project/workspace to scope to (defaults to this connection's workspace; isolates work streams within your account) |
Changes observed during successful MCP inspections. Dates show when Glama detected each change.
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the burden of behavioral disclosure. It does convey meaningful behavior: results are ranked via hybrid semantic + full-text matching and emphasize conceptual relevance. However, it does not explicitly state that the operation is read-only, what the return shape is, or any access/workspace side effects beyond the schema.
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?
Two tight sentences with no filler. The first sentence states the core purpose, and the second adds a differentiating behavioral detail. Information is front-loaded and 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?
The description is adequate for basic invocation: it names the artifact type, the query context, and the ranking behavior. But without annotations or an output schema, it leaves gaps around return format, result count semantics, and explicit relationship to alternatives like wyrm_search or wyrm_vault_get.
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 67%, with query and workspace already described. The tool description adds context that query should describe the current task and that ranking is semantic/full-text, which clarifies how the query is interpreted. The 'limit' parameter remains undocumented in both schema and description, so the agent must infer its meaning from the name.
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 action ('Recall relevant memory artifacts') and a clear resource domain ('lessons, patterns, reasoning traces'). It also distinguishes itself from simple keyword search by explicitly describing the hybrid semantic + full-text ranking behavior, helping separate it from sibling search tools like wyrm_search.
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 phrase 'for the current task' gives context for when to use it, and 'not just keyword matches' hints that keyword-only search might be an alternative. However, it never explicitly names a sibling or states when not to use this tool, leaving the routing decision partially to inference.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Add one secure layer between your agents and this server.
Multiple tools have heavily overlapping purposes: wyrm_all_quests vs wyrm_quest_list, wyrm_decision_downstream vs wyrm_decision_trace, and wyrm_recall vs wyrm_search are near-duplicates in behavior. wyrm_capture also overlaps with wyrm_remember, wyrm_truth_set, wyrm_quest_add, and wyrm_decided_because, making it hard for an agent to reliably pick the right tool.
Most names share the wyrm_ prefix and many follow a resource-action pattern like truth_set, quest_list, and failure_record. However, the pattern is broken by names like wyrm_all_quests, wyrm_decided_because, wyrm_context_build, wyrm_remember, and wyrm_capture, mixing noun-first, verb-first, and non-patterned names.
At 31 tools this is above the heavy threshold, and several tools appear redundant or mergeable, such as wyrm_all_quests/quest_list, decision_trace/downstream, capture/remember, and context_build/session_prime. The breadth is real, but the count feels inflated by overlapping entry points rather than distinct capabilities.
The tool surface covers core lifecycles well: quests have add/claim/update/complete/release, failures have record/check/list/resolve, decisions have record/trace/invalidate, and the vault has full get/put/list/delete. The main gaps are the lack of delete/update for memory artifacts and ground truths, and no explicit way to remove stale artifacts.
