core-mcp
REQUIRED before any wsh terminal operation. Contains the complete MCP tool reference and bootstrap sequence for wsh_create_session, wsh_send_input, wsh_get_screen, wsh_send_and_read, wsh_send_keys, and all wsh_* tools. Do NOT guess wsh CLI commands or HTTP endpoints — use MCP tools or load this skill first.
64
76%
Does it follow best practices?
Impact
—
No eval scenarios have been run
Passed
No known issues
Optimize this skill with Tessl
npx tessl skill review --optimize ./skills/core-mcp/SKILL.mdQuality
Discovery
75%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
The description is strong on completeness and distinctiveness, clearly stating when Claude must use this skill and carving out a unique niche around wsh MCP tools. However, it reads more as an internal directive than a capability description — it lists tool names without explaining what actions they enable (e.g., running shell commands, reading terminal output), and the trigger terms are technical rather than user-facing.
Suggestions
Add concrete action descriptions for the tools, e.g., 'Enables running shell commands, reading terminal output, and managing terminal sessions via wsh MCP tools.'
Include natural-language trigger terms a user might say, such as 'run a command', 'terminal session', 'shell', 'execute script', alongside the technical wsh_* tool names.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | It names the domain (wsh terminal operations) and lists specific tool names (wsh_create_session, wsh_send_input, etc.), but doesn't describe what concrete actions those tools perform — it's more of a tool inventory than a capability description. | 2 / 3 |
Completeness | It clearly answers 'what' (complete MCP tool reference and bootstrap sequence for wsh tools) and 'when' (REQUIRED before any wsh terminal operation; use MCP tools or load this skill first; do NOT guess wsh CLI commands). The trigger guidance is explicit and directive. | 3 / 3 |
Trigger Term Quality | It includes relevant keywords like 'wsh', 'terminal', 'MCP tool', and specific tool names like 'wsh_create_session', 'wsh_send_input'. However, these are technical/internal terms rather than natural language a user would say; users might say 'run a command', 'open a terminal', or 'shell session' instead. | 2 / 3 |
Distinctiveness Conflict Risk | The description is highly specific to wsh terminal MCP tools, with distinct tool names and a clear niche. It is unlikely to conflict with other skills due to the specificity of 'wsh_*' tools and the MCP tool reference context. | 3 / 3 |
Total | 10 / 12 Passed |
Implementation
77%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a strong, comprehensive MCP tool reference that provides highly actionable, concrete guidance with exact tool invocations and return values. Its main weakness is length — at 350+ lines it tries to be both a quick-start guide and a complete API reference in one file, which hurts conciseness and progressive disclosure. The workflow guidance is appropriate for a reference/bootstrap skill, with clear sequencing and good delegation to specialized sub-skills.
Suggestions
Split the detailed API sections (Visual Elements, Session Management, Federation) into separate reference files and link to them from SKILL.md, keeping the main file focused on the bootstrap sequence and fundamental loop.
Remove the metaphorical prose ('eyes/hands/patience/voice', 'This is your heartbeat. Learn it.') and the authentication section for local-only use to tighten token efficiency.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is fairly comprehensive and mostly efficient, but includes some unnecessary explanatory prose (e.g., the 'eyes/hands/patience/voice' metaphor, 'This is your heartbeat. Learn it.', explaining what a Unix domain socket is). Some sections like authentication could be trimmed since local use needs no auth. However, most content is reference material that earns its place. | 2 / 3 |
Actionability | Every MCP tool is documented with concrete, copy-paste-ready invocation examples showing exact parameter names and values. Named keys are enumerated, return values are specified with exact JSON shapes, and the getting-started sequence is fully executable step by step. | 3 / 3 |
Workflow Clarity | The 'Getting Started' section provides a clear 3-step sequence with a fallback path if MCP tools aren't available. The 'Fundamental Loop' (send → wait → read → decide) is explicitly defined. The skill appropriately delegates complex multi-step workflows to specialized sub-skills (drive-process, tui, etc.), and for a reference skill this level of workflow guidance is sufficient. | 3 / 3 |
Progressive Disclosure | The skill references specialized sub-skills (wsh:drive-process, wsh:tui, etc.) which is good progressive disclosure, and uses a collapsible details block for the JSON escape reference. However, the main file is very long (~350+ lines) and could benefit from splitting the detailed API reference (visual elements, session management, federation) into separate reference files, keeping SKILL.md as a concise overview with links. | 2 / 3 |
Total | 10 / 12 Passed |
Validation
90%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 10 / 11 Passed
Validation for skill structure
| Criteria | Description | Result |
|---|---|---|
frontmatter_unknown_keys | Unknown frontmatter key(s) found; consider removing or moving to metadata | Warning |
Total | 10 / 11 Passed | |
- Repository
- deepgram/wsh
- Commit
4863aaf
- Reviewed
Table of Contents
Is this your skill?
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.