large-workspace-handling
Rosetta skill to partition large workspaces or folders (100+ files recursively) into scoped subagent tasks when single-agent context is insufficient.
44
46%
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 ./instructions/r2/core/skills/large-workspace-handling/SKILL.mdQuality
Discovery
57%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 identifies a clear and distinctive niche—partitioning large workspaces into subagent tasks—but relies on technical jargon that users wouldn't naturally use and lacks an explicit 'Use when...' clause. It would benefit from more natural trigger terms and a clearer separation of what it does versus when it should be selected.
Suggestions
Add an explicit 'Use when...' clause, e.g., 'Use when the user needs to process a large codebase, split work across subagents, or when a single agent cannot handle the full file set in context.'
Include natural trigger terms users would say, such as 'too many files', 'split work', 'large codebase', 'parallelize tasks', 'divide project', rather than relying on internal jargon like 'Rosetta skill' and 'single-agent context'.
List additional concrete actions beyond partitioning, such as 'scans directory trees, estimates token budgets per partition, and assigns file groups to subagent tasks'.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | It names the domain (partitioning large workspaces) and one core action (partition into scoped subagent tasks), but doesn't list multiple concrete actions beyond that single capability. The mention of '100+ files recursively' adds some specificity. | 2 / 3 |
Completeness | The 'what' is reasonably clear (partition large workspaces into subagent tasks), but there is no explicit 'Use when...' clause. The 'when' is only implied by the condition 'when single-agent context is insufficient', which is not framed as explicit trigger guidance. | 2 / 3 |
Trigger Term Quality | Includes some relevant terms like 'large workspaces', 'folders', '100+ files', 'subagent tasks', but uses technical jargon ('Rosetta skill', 'scoped subagent tasks', 'single-agent context') that users wouldn't naturally say. Missing natural trigger terms like 'split work', 'too many files', 'parallelize', or 'divide and conquer'. | 2 / 3 |
Distinctiveness Conflict Risk | The description carves out a clear niche: partitioning large file sets into subagent tasks specifically when context limits are exceeded. This is unlikely to conflict with other skills due to the specific scope (100+ files, subagent decomposition). | 3 / 3 |
Total | 9 / 12 Passed |
Implementation
35%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill provides a reasonable conceptual framework for partitioning large workspaces into subagent tasks, with useful distinctions between summarization and work distribution strategies. However, it falls short on actionability—there are no concrete examples of subagent dispatch calls, scope partition templates, or executable workflows. The content reads more like a design document than an operational skill, leaving Claude to infer most implementation details.
Suggestions
Add a concrete, copy-paste-ready example showing how to partition a sample workspace (e.g., a monorepo with 3 packages) into scoped subagent tasks, including the exact subagent dispatch format with scope paths, goals, and boundaries.
Define an explicit numbered workflow for each strategy (Summarize & Index, Work distribution) with validation checkpoints—e.g., 'Step 3: Verify no file appears in more than one scope before dispatching subagents.'
Replace vague bullet points like 'Subagent: discoverer, explore, etc.' with a concrete subagent specification template showing required fields and example values.
Add a verification/feedback loop section detailing what to check after subagents complete, how to handle failures or scope conflicts, and when to re-dispatch.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The content is moderately efficient but includes some redundancy and vague bullet points that could be tightened. Phrases like 'Subagent: discoverer, explore, etc.' and 'Subagent: executor, engineer, etc.' are vague filler. The XML wrapper tags add unnecessary tokens. However, it doesn't over-explain basic concepts Claude already knows. | 2 / 3 |
Actionability | The skill provides abstract descriptions and bullet-point lists rather than concrete, executable guidance. There are no code examples, no specific commands, no example subagent dispatch calls, and no template for how to actually partition a workspace. Instructions like 'Spawn subagents in parallel if possible' and 'Resolve shared-interface conflicts or changes with extra pass' are vague directives without concrete implementation details. | 1 / 3 |
Workflow Clarity | There is an implicit sequence (prep → scope → dispatch → verify), and the two strategies are distinguished. However, the steps lack explicit ordering, validation checkpoints are minimal (only a brief mention of spawning verification subagents at the end), and there are no feedback loops for error recovery. The verification step is mentioned but not detailed. | 2 / 3 |
Progressive Disclosure | The content references external skills (reverse-engineering/SKILL.md, init-workspace-discovery/SKILL.md, load-context) which shows some progressive disclosure. However, with no bundle files provided, the references can't be verified. The content itself is somewhat monolithic—the two strategies and scoping rules could benefit from clearer section separation or external reference files for detailed subagent templates. | 2 / 3 |
Total | 7 / 12 Passed |
Validation
90%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 10 / 11 Passed
Validation for skill structure
| Criteria | Description | Result |
|---|---|---|
frontmatter_unknown_keys | Unknown frontmatter key(s) found; consider removing or moving to metadata | Warning |
Total | 10 / 11 Passed | |
- Repository
- griddynamics/rosetta
- Commit
ade2cc5
- 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.