knowledge
Extract, search, and retrieve knowledge from Rosetta reference documentation. Use when the user asks for Rosetta system knowledge, wants to understand specific components (agents, skills, rules, workflows, templates), needs setup/installation help, or requires guidance on bootstrap policies, architecture, patterns, or operations. Handles topic searches, document navigation, and contextual knowledge retrieval.
56
63%
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 ./.claude/skills/knowledge/SKILL.mdQuality
Discovery
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.
This is a well-crafted skill description that clearly defines its scope around Rosetta reference documentation. It excels in all dimensions: it lists concrete actions, provides abundant natural trigger terms, includes an explicit 'Use when' clause with multiple trigger scenarios, and is highly distinctive due to its domain-specific terminology. The description is concise yet comprehensive.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Lists multiple specific concrete actions: 'Extract, search, and retrieve knowledge', 'topic searches, document navigation, and contextual knowledge retrieval'. Also enumerates specific components like agents, skills, rules, workflows, templates, bootstrap policies, architecture, patterns, and operations. | 3 / 3 |
Completeness | Clearly answers both 'what' (extract, search, retrieve knowledge from Rosetta reference documentation; handles topic searches, document navigation, contextual knowledge retrieval) and 'when' (explicit 'Use when' clause covering multiple trigger scenarios: asking for system knowledge, understanding components, setup/installation help, guidance on architecture/patterns). | 3 / 3 |
Trigger Term Quality | Includes strong natural trigger terms users would say: 'Rosetta', 'agents', 'skills', 'rules', 'workflows', 'templates', 'setup', 'installation', 'bootstrap policies', 'architecture', 'patterns', 'operations'. These cover a wide range of natural user queries about the Rosetta system. | 3 / 3 |
Distinctiveness Conflict Risk | Highly distinctive due to the specific domain anchor 'Rosetta reference documentation' and the enumeration of Rosetta-specific concepts (bootstrap policies, agents, skills, rules, workflows). Unlikely to conflict with other skills unless there are multiple Rosetta-related skills. | 3 / 3 |
Total | 12 / 12 Passed |
Implementation
27%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill is a minimal, high-level outline that lacks the concrete, actionable detail needed for reliable execution. It describes a workflow at an abstract level without providing executable examples, subagent invocation syntax, expected output formats, or validation steps. The concept is sound but the implementation guidance is insufficient.
Suggestions
Add a concrete example showing how to invoke a discoverer subagent, including the exact tool call syntax, model parameter, task description format, and expected response structure.
Include an example end-to-end workflow: a sample user query, the INDEX.md lookup step, the subagent delegation messages, and the final synthesized answer format.
Add validation/quality checkpoints: how to verify subagent responses are relevant, what to do if a subagent returns insufficient information, and criteria for when to read a referenced file directly.
Structure the content with clear sections (e.g., ## Workflow, ## Subagent Delegation, ## Synthesis) and provide links or references to any supporting files like INDEX.md format documentation.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Reasonably brief but includes some vague/unnecessary lines like 'you must follow delegation best practices' and 'You should ask slightly more from each subagent vs what you have been requested' which are imprecise and add little actionable value. | 2 / 3 |
Actionability | The skill provides only high-level abstract instructions with no concrete code, commands, subagent invocation syntax, example queries, or expected output formats. 'Spawn parallel discoverer subagents with fast/haiku/flash/mini model' lacks any executable detail on how to actually do this. | 1 / 3 |
Workflow Clarity | There is a rough sequence (read INDEX.md → don't read files → spawn subagents → synthesize), but there are no validation checkpoints, no error handling, no feedback loops for when subagents return incomplete or conflicting information, and no explicit criteria for what constitutes a sufficient answer. | 2 / 3 |
Progressive Disclosure | References `refsrc/INDEX.md` but no bundle files are provided, no navigation structure exists, and there are no links to supporting documentation. The skill is a flat list of bullets with no structural organization or signposted references. | 1 / 3 |
Total | 6 / 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
- griddynamics/rosetta
- Commit
0f3002b
- 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.