new-boost-module

Create new Harbor Boost modules — the Python plugins that run inside Harbor's LLM proxy. Use this skill whenever the user wants to build a Boost module, write a custom module for Harbor Boost, add a new feature to the Boost proxy pipeline, or create any kind of middleware that transforms, augments, or intercepts LLM chat completions in Harbor. Also triggers when the user mentions "boost module", "boost plugin", "custom module for boost", or wants to add prompt engineering, reasoning chains, or output transforms to Harbor's proxy layer.

Quality

85%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

75%Weight 40%Scale 1-5

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

A strong, code-heavy skill body that teaches the module structure, primitives, and common patterns with executable examples and a completion checklist. It falls just short of top marks on all dimensions: imports for `ch`/`log` are never shown, there is no run-and-verify step, and the inlined API reference plus the long DOCS example could be slimmed or split out.

Suggestions

Add a short 'Test your module' section to the workflow (e.g., restart Boost and invoke the model `mymod-llama3.1` to confirm the module loads and behaves) so the checklist has a runtime validation checkpoint.

Show the imports behind the snippets — e.g., how `ch` (ChatNode/Chat) and `log` are obtained inside a module — so the copy-paste examples are fully executable.

Move the chat/llm primitive API listings into a references/ file (linked from a brief summary) and trim the DOCS example's docker run block to keep SKILL.md a lean overview.

Dimension	Reasoning	Score
Conciseness	Mostly efficient — lean tables and copy-ready code with no padding about concepts Claude already knows — but the full DOCS format example (including the entire docker run block) and the motivational line about primitives 'saving you from reinventing patterns' could be trimmed.	4 / 5
Actionability	Concrete, mostly copy-paste-ready code for the module skeleton, every primitive, and three common patterns. Minor gaps keep it below a 5: snippets use `ch.ChatNode`/`ch.Chat` and `logger`/`log.setup_logger` without ever showing the corresponding imports.	4 / 5
Workflow Clarity	A clear sequence exists ('Before You Write Code' 1–3, then structure → primitives → patterns) and the final checklist acts as a verification checkpoint, but there is no step validating that the module actually works (e.g., invoking `mymod-llama3.1` and checking the output).	4 / 5
Progressive Disclosure	No bundle files exist; the single SKILL.md is well-sectioned with clear pointers to repo docs (docs/5.2.1, docs/5.2.3) and existing modules. The ~80 lines of inlined chat/llm API reference could arguably move to a references file, which keeps this at 'good structure with minor organization gaps' rather than 5.	4 / 5
	Total	16 / 20 Passed

Description

96%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

An excellent description: it states what the skill does concretely, gives explicit 'Use when...' trigger guidance with quoted natural phrases, and stays in third-person voice. The only weakness is slightly broad middleware phrasing that could marginally overlap with a general LLM-proxy skill.

Dimension	Reasoning	Score
Specificity	Lists multiple specific concrete actions — 'build a Boost module', 'write a custom module for Harbor Boost', 'add a new feature to the Boost proxy pipeline', 'create any kind of middleware that transforms, augments, or intercepts LLM chat completions' — with comprehensive coverage of the skill's scope.	5 / 5
Completeness	Clearly answers both parts: the 'what' ('Create new Harbor Boost modules — the Python plugins that run inside Harbor's LLM proxy') and an explicit 'when' ('Use this skill whenever the user wants to... Also triggers when the user mentions...') with concrete trigger phrases.	5 / 5
Trigger Term Quality	Explicitly enumerates the natural phrases a user would say — "boost module", "boost plugin", "custom module for boost" — plus variations like 'prompt engineering', 'reasoning chains', and 'output transforms', giving comprehensive synonym coverage.	5 / 5
Distinctiveness Conflict Risk	The 'boost module/plugin' triggers form a clear niche with minimal conflict risk, but the broad phrasing 'create any kind of middleware that transforms, augments, or intercepts LLM chat completions' leaves minor overlap risk with a hypothetical general LLM-proxy/middleware skill.	4 / 5
	Total	19 / 20 Passed

Validation

100%

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 — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: av/harbor
Path: .agents/skills/new-boost-module/SKILL.md
Commit: e38ae4c

Reviewed: 1 day 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.