Content
67%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A well-structured router whose routing table and resources section provide genuinely actionable, well-signaled navigation to per-tool references. Its main weaknesses are the large inlined generic routing-discipline block and the Device Hub duplication, which bloat the token budget, plus references that point to files outside the skill's own bundle.
Suggestions
Move the 'Using Axiom Skills' session-start block into a shared reference file and keep a one-line pointer, cutting roughly 90 lines of inlined generic routing rules.
Trim the Device Hub paragraph to a two-sentence summary plus the existing pointer, since the subcommand matrix and status-bar details already live in skills/device-control-ref.md.
Add one concrete example invocation per tool (e.g. a sample 'xclog' capture command) in the Resources list to make the routing immediately executable.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The routing table and Resources section are lean, but the ~95-line inlined 'Using Axiom Skills' block pads with emphatic repetition ('This is not negotiable. This is not optional. You cannot rationalize your way out of this.') and a rationalization table, and the Device Hub paragraph restates details that the pointer says live in skills/device-control-ref.md. Mostly efficient with sections that should be tightened, matching anchor 3; not anchor 2 because the core routing content is dense and earns its tokens. | 3 / 5 |
Actionability | Concrete, executable guidance throughout: a fallback path rule ('use `<plugin-root>/tools/<tool>` in place of the bare name'), a question-to-reference routing table with a specific command ('xcui resize sweep'), and named per-tool pointers. Minor gaps — most tools get no example invocation and the `<plugin-root>` resolution rule is convoluted — keep it at anchor 4 rather than 5. | 4 / 5 |
Workflow Clarity | The router's decision flow is clear: check tool availability, route the question via the table to a specific file/section, and the priority ordering (environment → architecture → implementation) plus the multi-domain rule ('invoke ALL relevant routers') sequence decisions well. No validation checkpoints exist, but this router performs no destructive or batch operations, so the cap does not apply; it stays at 4 because the overall read-then-route sequence is implicit rather than stated as steps. | 4 / 5 |
Progressive Disclosure | Good structure: an overview with a routing table mapping questions to one-level-deep, clearly signaled references (skills/xclog-ref.md, skills/xcsym-ref.md, etc.). It falls short of anchor 5 because the ~95-line generic session-start block is inlined rather than kept in its own file, and the referenced skills/*-ref.md files are not present in this skill's bundle (no references/ directory), leaving some pointers unverifiable. | 4 / 5 |
Total | 15 / 20 Passed |