Content
67%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a well-structured orchestration skill that clearly defines a phased workflow for helping users shape their work into reusable context. Its strengths are in workflow sequencing, clear boundaries (especially the eval separation), and principled guidance. Its main weakness is the lack of concrete examples — no sample inputs/outputs, no example CLI invocations with real arguments, and no illustration of what 'well-shaped context' looks like in practice.
Suggestions
Add a brief concrete example showing what a typical user request looks like and how the arc plays out (e.g., 'User says X → gather phase produces Y → plan phase produces Z').
Include at least one complete CLI command example with arguments (e.g., `tessl plugin new my-plugin --description '...'`) to make the actionability more concrete.
Add explicit validation gates between phases, such as 'Before moving to phase 3, confirm the plan with the user and verify all referenced artifacts exist.'
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The content is mostly efficient and respects Claude's intelligence. It avoids explaining what plugins or skills are in basic terms. Minor verbosity in the 'Principles' section (e.g., 'A plugin must earn its place' is somewhat philosophical) and the 'Then, separately: eval' section repeats the 'don't run evals here' point already made in the intro. | 4 / 5 |
Actionability | The skill provides a clear arc with named sub-skills to invoke (gather-context, plan-composition, build-composition, etc.) and mentions specific CLI commands (tessl skill new, tessl plugin new). However, it lacks concrete examples of inputs/outputs, specific command invocations with arguments, or executable code. The guidance is structural rather than executable — it tells Claude which phases to run and which sub-skills to call, but the actual 'how' is delegated entirely to those sub-skills without any inline examples. | 3 / 5 |
Workflow Clarity | The four-phase arc is clearly sequenced and numbered, with explicit phase names and sub-skill references. There's a clear 'when to stop' section and the eval handoff is well-defined as a boundary. However, there are no explicit validation checkpoints between phases (e.g., 'confirm you have enough material before moving to phase 2' is implied in phase 1's description but not formalized as a gate), and no error recovery/feedback loops are specified. | 4 / 5 |
Progressive Disclosure | The skill effectively delegates to sub-skills (gather-context, plan-composition, build-composition, decompose-into-skills, publish-plugin) and references a sibling flow (tessl/skill-optimizer) for eval. The structure is clean with well-organized sections. However, no bundle files are provided to verify these references resolve correctly, and the references to sub-skills are inline text rather than linked paths, making navigation slightly less clear. | 4 / 5 |
Total | 15 / 20 Passed |