BA SaaS
Runs a SaaS business: subscription revenue, plan packaging, trials, retention, expansion, and enterprise readiness. Use when MRR or ARR has to be computed, reconciled, or explained; when NRR, gross margin, CAC payback, burn multiple, or rule of 40 is the question; when designing plans, seats, usage limits, add-ons, or a free tier; when trials sign up but never convert; when failed payments leak revenue; when a renewal, downgrade, or cancellation flow needs building; when expansion has stalled; when an enterprise buyer demands SSO, SCIM, audit logs, SOC 2, a DPA, or an uptime SLA; when per-tenant cost or AI inference COGS breaks the margin; when designing multi-tenant isolation, metering, or entitlements; or when an investor questions ARR quality. Not for setting the price itself (`pricing`), acquisition channels and funnel diagnosis (`growth`), building the payment integration (`billing`), company financial models (`cfo`), or closing an individual deal (`b2b`).
As a process A 80/100 · Runs to the end — weak spots: inputs and preconditions, progress reporting
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 · 5
✓ No critical or high findings
Medium and low: 5
-
low Risky intent
intent-offensive-securitycompliance.md:73Offensive-security / dual-use content (legitimate for authorised testing; review intended use)**After any compliance decision or milestone**, write the certification, its scope and its dates to `## Commitments` where a customer was promised it, add evidence collection, audit renewal, penetrati
-
low Risky intent
intent-offensive-securitydiligence.md:41Offensive-security / dual-use content (legitimate for authorised testing; review intended use)| Security and compliance | SOC 2 or ISO report, penetration test summary, subprocessor list, security answer bank (`compliance.md`) |
-
low Risky intent
intent-offensive-securitymemory-template.md:140Offensive-security / dual-use content (legitimate for authorised testing; review intended use)| Penetration test | year | 2025-11-10 | 2026-11-10 |
-
low Risky intent
intent-offensive-securitymemory-template.md:285Offensive-security / dual-use content (legitimate for authorised testing; review intended use)- Contains no credentials, no architecture diagrams with internal hostnames, and no findings from an unremediated penetration test.
-
low Risky intent
intent-offensive-securitySKILL.md:171Offensive-security / dual-use content (legitimate for authorised testing; review intended use)- **Cadence** — revenue close day, board and investor update frequency, renewal notice window, pricing review, audit and penetration test, dunning review — every accepted cadence becomes a row in the
Files scanned: 21. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- warning
name-formatname should be kebab-case (lowercase letters, digits, hyphens) - warning
body-longSKILL.md body ≈ 7272 tokens (recommended < 5000); move details to references/ - note
frontmatter-keyunknown frontmatter key "slug" - note
frontmatter-keyunknown frontmatter key "homepage" - note
frontmatter-keyunknown frontmatter key "changelog"
Process rating: all ten parameters 80/100
- 0Inputs and preconditions. Does not say what the process needs to start
- 0Progress reporting. Says nothing while it works
- 60Result and completion. Output format stated, no completion criterion
- 70Execution cost. Instruction body is 7272 tokens
- 100Tools and files. No external tools needed
- 100Steps. 44 steps
- 100When it triggers. States when to use and when not to
- 100Failures and branches. 5 branches, has a failure section
- 100Consistency. Name and required fields are in place
- 100Running it twice. Mutating operations check current state
- medium Safety rules and hard prohibitions inside a skill: they belong in the system prompt, here they protect nothing
- low 13 top-level sections: this looks like several domains in one skill
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
- +3Description length 975: 120–800 characters recommended
- +4No input/output examples
- +1No license
- +2Single-language instructions
- +4Description says when NOT to use the skill
- +4Structure: 13 headings
- +3Step-by-step instructions: 44 items
- +3Output format is stated explicitly
Quality base 70; lint remarks subtract, signals add up to 100. Result: 66.