Skill Retriever
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tasks | {
"list": {},
"cancel": {},
"requests": {
"tools": {
"call": {}
},
"prompts": {
"get": {}
},
"resources": {
"read": {}
}
}
} |
| tools | {
"listChanged": true
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| search_componentsC | Search components by task. |
| get_component_detailC | Get full component info. |
| install_componentsB | Install components to .claude/. |
| check_dependenciesD | Check deps and conflicts. |
| ingest_repoB | Index a component repository. |
| register_repoC | Register a repo for auto-sync. |
| unregister_repoB | Unregister a repo from auto-sync. |
| list_tracked_reposB | List all tracked repos. |
| sync_statusB | Get sync system status. |
| start_sync_serverB | Start webhook server and poller. |
| stop_sync_serverB | Stop webhook server and poller. |
| poll_repos_nowB | Trigger immediate poll of all repos. |
| run_discovery_pipelineD | Run the discovery and ingestion pipeline. |
| discover_reposB | Discover skill repositories from GitHub. |
| get_heal_statusB | Get auto-heal status and failures. |
| get_pipeline_statusB | Get discovery pipeline status. |
| clear_heal_failuresB | Clear all tracked failures from auto-heal. |
| report_outcomeC | Report a component outcome (used, removed, deprecated). |
| get_outcome_statsC | Get outcome statistics for a component. |
| get_outcome_reportC | Get overall outcome report with problematic components. |
| analyze_feedbackC | Analyze usage patterns and generate edge suggestions. |
| get_feedback_suggestionsB | Get pending edge suggestions from feedback analysis. |
| review_suggestionC | Review a pending edge suggestion (accept or reject). |
| apply_feedback_suggestionsB | Apply all accepted suggestions to the graph. |
| security_scanB | Scan a component for security vulnerabilities. |
| security_auditC | Audit all indexed components for security vulnerabilities. |
| backfill_security_scansB | Backfill security scans for existing components. Scans all components in the metadata store that don't have security data, or all components if force_rescan=True. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 27 tools
Most tools have distinct purposes, but some overlap exists that could cause confusion. For example, 'security_scan' and 'security_audit' both handle security checks, and 'run_discovery_pipeline' overlaps with 'discover_repos' and 'ingest_repo'. Descriptions help clarify, but agents might misselect between related tools.
Tool names follow a consistent snake_case verb_noun pattern throughout, such as 'analyze_feedback' and 'get_component_detail'. There are minor deviations like 'backfill_security_scans' using plural 'scans' while others use singular nouns, but overall the naming is predictable and readable.
With 27 tools, the count is too high for the server's apparent scope of skill retrieval and management. This many tools suggests fragmentation or redundancy, making it heavy for agents to navigate and likely including overlapping functionalities that could be consolidated.
The tool set covers a broad range of operations for skill retrieval, including discovery, ingestion, security, feedback, and sync management. Minor gaps exist, such as no explicit tool for updating component metadata or handling errors in detail, but core workflows are well-supported with CRUD-like operations.