agents-connect

Use when connecting your agent to external APIs, tools, or services via Gateway, or restricting tool access with Cedar policies. Handles gateway setup, target types, outbound auth (OAuth, API key, IAM), credentials, and Cedar policy authoring. Triggers on: "connect to API", "add gateway", "connect to MCP server", "Lambda tools", "OpenAPI", "gateway target", "Cedar policy", "restrict tools", "policy engine", "gateway auth error", "store API key", "outbound credential", "env var API key", "API key None after deploy", "credential not available after deploy", "should this be a gateway target", "give my agent tools", "add tools to agent". Not for inbound auth (who can call your agent) — use agents-harden. Not for debugging agent behavior — use agents-debug. Not for VPC networking errors (agent can't reach APIs due to VPC) — use agents-build. Not for creating or hosting a new MCP server project — use agents-get-started.

Quality

88%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

High

Do not use without reviewing

Quality

Content

77%

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

The body is highly actionable with executable commands and complete code, and the workflow is well-sequenced with useful gates. Its weaknesses are verbosity from cross-section repetition and a monolithic inline structure that underuses progressive disclosure.

Suggestions

Dedupe the 3LO/authorization-code and MCP-api-key-not-supported explanations: state each once in the auth matrix and have Path A / Troubleshooting link back rather than re-explain.

Move the per-framework gateway client code (Strands, LangGraph) and/or the full Troubleshooting section into reference files, keeping SKILL.md as an overview with one-level-deep links.

Tighten conceptual framing sections (e.g., "Before jumping into paths, set expectations" and the mental-model walkthroughs) to the minimum needed to steer the triage decision.

Dimension	Reasoning	Score
Conciseness	The ~555-line body is mostly high-signal AgentCore specifics, but repeats key facts across sections (3LO explained in the auth matrix, Path A, and Troubleshooting; MCP-api-key-unsupported repeated three times; the five direct-call cases listed twice) and includes framing prose that could be trimmed.	2 / 3
Actionability	Provides fully executable `agentcore` CLI commands, complete copy-paste Python for the framework-agnostic MCP client, Strands, LangGraph, and the @requires_api_key/@requires_access_token decorators, plus a concrete tools.json schema — copy-paste ready.	3 / 3
Workflow Clarity	Sequenced Steps 0–3 with explicit gates (CLI version check at Step 0, the "Before you reach for Path D" confirmation gate, env-var verification before using the gateway URL) and a troubleshooting section for error recovery, matching the clear-sequence-with-checkpoints anchor.	3 / 3
Progressive Disclosure	The single `references/policy.md` reference is real and clearly signaled one level deep, but the large body keeps most detail inline (framework-specific code, auth matrix, troubleshooting) rather than splitting it into reference files, so content is not appropriately split.	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, third-person, and rich with natural trigger terms, while explicitly bounding its scope against four sibling skills. It cleanly answers both what the skill does and when to use it.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete capabilities — "Handles gateway setup, target types, outbound auth (OAuth, API key, IAM), credentials, and Cedar policy authoring" — matching the anchor for several specific concrete actions.	3 / 3
Completeness	Explicitly answers both what (connecting agents via Gateway, restricting with Cedar) and when via an explicit "Use when..." clause plus a "Triggers on:" list, satisfying the anchor for clear what-and-when with explicit triggers.	3 / 3
Trigger Term Quality	The "Triggers on:" list covers many natural phrases a user would actually say ("connect to API", "add gateway", "give my agent tools", "API key None after deploy"), giving strong coverage of natural terms.	3 / 3
Distinctiveness Conflict Risk	Actively disambiguates from sibling skills ("Not for inbound auth — use agents-harden", "Not for debugging agent behavior — use agents-debug", etc.), giving a clear niche unlikely to trigger the wrong skill.	3 / 3
	Total	12 / 12 Passed

Validation

87%

Warnings & errors only

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

Validation — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (557 lines); consider splitting into references/ and linking	Warning
relative_links	Relative link issues: 4 suspicious	Warning

	Total	14 / 16 Passed

Repository: aws/agent-toolkit-for-aws
Path: plugins/aws-agents/skills/agents-connect/SKILL.md
Commit: 222ce56

Reviewed: 3 days ago

Table of Contents

Discovery Implementation Validation

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.