AB deepen
Point at any topic to build a compounding, domain-general expertise knowledge base: a high-fidelity source corpus, track-record-weighted experts, reconciled falsifiable stances, failure modes, benchmarks, and a proven playbook — built as evergreen notes a future agent can load and reason over. Use when the user wants to deeply master, research, or build durable, defensible knowledge on a topic — not for a quick one-off answer or a single throwaway report.
Point at any topic to build a compounding, domain-general expertise knowledge base: a high-fidelity source corpus, track-record-weighted experts, reconciled…
As a process B 72/100 · Nearly there — weak spots: inputs and preconditions
How to improve
- The SKILL.md body is over 5,000 tokens: move reference detail into references/ and load it when needed.
- Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
- A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.
Guard findings · 1
✓ No critical or high findings
Medium and low: 1
-
low Concealment
en-hide-from-userSKILL.md:181Instruction to hide actions from the user (negated — the text forbids it)**Supersede rule:** when new evidence contradicts an existing note, *replace the claim in place* and log `superseded: <old> → <new> (why)` in `changelog.md`. Never leave conflicting claims live; never
negated
Files scanned: 4. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- warning
body-longSKILL.md body ≈ 7870 tokens (recommended < 5000); move details to references/
Process rating: all ten parameters 72/100
- 0Inputs and preconditions. Does not say what the process needs to start
- 60Tools and files. Uses tools (bash, web) that frontmatter does not declare
- 60Result and completion. Output format stated, no completion criterion
- 70Execution cost. Instruction body is 7870 tokens
- 85Steps. 70 steps, 1 vague phrases
- 100When it triggers. States when to use and when not to
- 100Failures and branches. 2 branches, has a failure section
- 100Consistency. Name and required fields are in place
- 100Running it twice. Mutating operations check current state
- 100Progress reporting. Reports progress
- medium Safety rules and hard prohibitions inside a skill: they belong in the system prompt, here they protect nothing
- high The skill tells the model to perform an irreversible action with no human approval
- low The response is described with custom markup (12 tags): a typed call is more reliable
Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.
Quality signals
- +5Description has no quoted example phrases that should trigger the skill
- +4No input/output examples
- +2Single-language instructions
- +4Description says when NOT to use the skill
- +3Description length 459: enough signal without eating the budget
- +4Structure: 27 headings
- +3Step-by-step instructions: 70 items
- +3Output format is stated explicitly
- +1License stated
Quality base 70; lint remarks subtract, signals add up to 100. Result: 78.