create-connector
Generate Harness Connector YAML for integrations and create/test via MCP. Use when user says "create connector", "git connector", "aws connector", "docker connector", "cloud connector", or wants to connect Harness to external services.
83
73%
Does it follow best practices?
Impact
100%
1.13xAverage score across 3 eval scenarios
Passed
No known issues
Optimize this skill with Tessl
npx tessl skill review --optimize ./plugins/claude/skills/create-connector/SKILL.mdQuality
Discovery
89%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This is a solid skill description that clearly defines its niche (Harness Connector YAML generation) and provides explicit trigger guidance with multiple natural keyword variations. The main weakness is that the 'what' portion could be more specific about the range of concrete actions supported beyond generate/create/test. Overall, it performs well across all dimensions.
Suggestions
Expand the specificity of capabilities by listing more concrete actions, e.g., 'Generate Harness Connector YAML, validate connector configurations, test connectivity, and manage connector settings via MCP.'
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | The description names the domain (Harness Connector YAML) and mentions some actions ('Generate', 'create/test via MCP'), but doesn't list multiple specific concrete actions like listing connector types, validating configurations, or updating existing connectors. | 2 / 3 |
Completeness | Clearly answers both 'what' (Generate Harness Connector YAML for integrations and create/test via MCP) and 'when' (explicit 'Use when' clause with specific trigger phrases and a general condition about connecting to external services). | 3 / 3 |
Trigger Term Quality | Includes strong natural trigger terms users would actually say: 'create connector', 'git connector', 'aws connector', 'docker connector', 'cloud connector', and 'connect Harness to external services'. Good coverage of common variations. | 3 / 3 |
Distinctiveness Conflict Risk | Highly specific to Harness Connector YAML generation, which is a clear niche. The combination of 'Harness', 'Connector', 'YAML', and 'MCP' makes it very unlikely to conflict with other skills. | 3 / 3 |
Total | 11 / 12 Passed |
Implementation
57%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The skill excels at actionability with comprehensive, copy-paste-ready YAML templates and concrete MCP tool invocations. However, it suffers from being a monolithic document that dumps all connector types inline, making it very long and token-expensive. The workflow is reasonable but lacks explicit feedback loops for error recovery after failed connection tests.
Suggestions
Split connector templates into separate reference files (e.g., git-connectors.md, cloud-connectors.md, registry-connectors.md) and reference them from the main SKILL.md to improve progressive disclosure and reduce token cost.
Consolidate repetitive YAML patterns—many connectors share identical UsernamePassword auth blocks that could be shown once with a note about reuse across types.
Add an explicit feedback loop to the workflow: 'If test_connection fails → check troubleshooting section → fix credentials/config → re-create connector → test again'.
Remove the 'Examples' section at the bottom as it merely restates what the templates already demonstrate, saving tokens.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is largely template-driven which is useful, but the sheer volume of YAML templates (GitHub, GitLab, Bitbucket, AWS, GCP, Azure, Docker Hub, ECR, K8s, Helm HTTP, Helm OCI, Nexus, Artifactory, Jira, ServiceNow) makes it very long. Many templates are structurally repetitive (e.g., all UsernamePassword auth patterns). The 'Examples' section at the end just restates what's already obvious from the templates. Some trimming and consolidation would improve token efficiency. | 2 / 3 |
Actionability | The skill provides fully formed, copy-paste-ready YAML templates for each connector type with inline comments explaining options. The MCP tool calls are concrete with specific parameter names. The troubleshooting section gives specific, actionable fixes (e.g., secret reference format `account.secret_name`). | 3 / 3 |
Workflow Clarity | The 4-step workflow at the top is clear and includes a test step. However, the validation checkpoint between creation and testing is implicit rather than explicit—there's no 'if test fails, fix and retry' feedback loop in the main workflow. The Performance Notes mention testing but don't integrate it into a structured feedback loop. For an operation that creates infrastructure resources, a more explicit validate-fix-retry cycle would be warranted. | 2 / 3 |
Progressive Disclosure | This is a monolithic wall of YAML templates with no external references or file splitting. At ~250+ lines, the connector templates for 15+ types should be split into separate reference files (e.g., git-connectors.md, cloud-connectors.md) with the SKILL.md serving as an overview. There are no bundle files to support progressive disclosure. | 1 / 3 |
Total | 8 / 12 Passed |
Validation
100%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 11 / 11 Passed
Validation for skill structure
No warnings or errors.
- Repository
- harness/harness-ai
- Commit
a000afa
- 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.