CtrlK
BlogDocsLog inGet started
Tessl Logo

external-agents

Connect external agents and MCP hosts (Claude, Claude Desktop, Claude Code, ChatGPT custom MCP apps, Codex, Cursor, Claude Cowork, VS Code GitHub Copilot, Goose, Postman, MCPJam) to an agent-native app over MCP, and round-trip artifacts back into the UI with MCP Apps and deep links. Use when adding an action's `link` builder or `mcpApp`, wiring the `/_agent-native/open` route, exposing an "ingest" action to MCP/A2A, or scaffolding apps from an external agent.

69

Quality

88%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Low

Low-risk findings worth noting

SKILL.md
Quality
Evals
Security

Quality

Content

77%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A highly actionable, well-sequenced body with executable code and built-in validation feedback loops, held back slightly by repetition across sections and a large amount of reference-grade content kept inline rather than progressively disclosed to bundle files.

Suggestions

Dedupe the host/client enumeration and Dispatch URL — define the client list and connector URL once in the Rule, then reference them from How/Do instead of repeating.

Move the verbose builtin-verb catalog (list_apps/open_app/ask_app/etc.) and per-host setup detail into a references file to lift progressive_disclosure and shrink the inline body.

Tighten the compact-catalog/auth-policy paragraph (section 1a) into a short rule plus a reference, since it currently re-explains the same policy across several sentences.

DimensionReasoningScore

Conciseness

The body is dense and assumes Claude's competence (no primers on MCP/OAuth), but the client enumeration and Dispatch URL recur across the Rule, How, and Do sections, and ~510 lines could be tightened — placing it at 'mostly efficient but could be tightened' rather than fully lean.

2 / 3

Actionability

Provides fully executable, copy-paste-ready guidance: real `buildDeepLink`/`embedApp`/`callMcpTool` code, exact `npx @agent-native/core@latest connect` commands, and precise config paths (.mcp.json, ~/.claude.json, ~/.codex/config.toml).

3 / 3

Workflow Clarity

The numbered How section (1, 1a, 1b, 2, 2a, 3, 4, 5) sequences the work and embeds explicit validation/feedback loops — "validate with a fresh tool call", reconnect/rescan on stale descriptors, sparse `tools/list` as a diagnostic, and test-after-shell-change guidance.

3 / 3

Progressive Disclosure

One reference is clearly signaled and real ("**`references/mcp-apps-embedding.md`**. Read it before changing..."), but the ~510-line body keeps large reference-grade material (verb catalog, per-host setup, connector internals) inline rather than split into the single available bundle file.

2 / 3

Total

10

/

12

Passed

Description

100%

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 specific, complete, and distinctive: it states concrete capabilities and provides an explicit 'Use when' trigger clause covering the main integration scenarios, all in third-person voice.

DimensionReasoningScore

Specificity

Lists multiple concrete actions — "Connect external agents and MCP hosts", "round-trip artifacts back into the UI with MCP Apps and deep links", "adding an action's `link` builder", "wiring the `/_agent-native/open` route", "exposing an 'ingest' action", "scaffolding apps" — matching the 'lists multiple specific concrete actions' anchor.

3 / 3

Completeness

Clearly answers both 'what' (connect external agents/MCP hosts over MCP, round-trip artifacts) and 'when' via an explicit "Use when..." clause with concrete triggers, satisfying the top anchor.

3 / 3

Trigger Term Quality

The explicit "Use when adding an action's `link` builder or `mcpApp`, wiring the `/_agent-native/open` route, exposing an 'ingest' action ... or scaffolding apps" clause plus enumerated client names (Claude, ChatGPT, Cursor, Codex) gives good coverage of natural terms a user of this skill would say.

3 / 3

Distinctiveness Conflict Risk

A clear niche — agent-native MCP bridging with deep links/MCP Apps — with distinct triggers (`link` builder, `mcpApp`, `/_agent-native/open`, ingest action) and third-person voice, making it unlikely to fire for the wrong skill.

3 / 3

Total

12

/

12

Passed

Validation

68%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation11 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

skill_md_line_count

SKILL.md is long (530 lines); consider splitting into references/ and linking

Warning

metadata_version

'metadata.version' is missing

Warning

metadata_field

'metadata' should map string keys to string values

Warning

frontmatter_unknown_keys

Unknown frontmatter key(s) found; consider removing or moving to metadata

Warning

relative_links

Relative link issues: 1 missing

Warning

Total

11

/

16

Passed

Repository
BuilderIO/agent-native
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.