AC architecture-critic
Adversarial pre-build architecture review. Spawns a structurally independent agent that reviews proposed builds, features, or operational decisions BEFORE any code is written. The critic sees only the task spec and codebase state — never conversation history, enthusiasm, or the proposer's reasoning. Returns APPROVE, REVISE, or REJECT with specific, itemized findings. Use when: starting any new feature, endpoint, schema change, payment/auth touch, or any decision being recommended with confidence. Skip only for: copy/style-only changes under 3 files with no logic, or isolated bug fixes that don't touch payment or auth.
As a process C 53/100 · Has gaps — weak spots: result and completion, when it triggers, inputs and preconditions
How to improve
- 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 · 2
✓ No critical or high findings
Medium and low: 2
-
low Instruction override
en-ignore-previousreferences/ai-builds.md:31Instruction-override phrase ("ignore previous instructions") (negated — the text forbids it)- [ ] **System prompt sealed** — instructions in system prompt; user input cannot override the system prompt via injection attempts
negated -
low Risky intent
intent-offensive-securityreferences/checklist-general.md:28Offensive-security / dual-use content (legitimate for authorised testing; review intended use)- Does this introduce new privilege escalation paths?
Files scanned: 11. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- note
frontmatter-keyunknown frontmatter key "requires" - note
frontmatter-keyunknown frontmatter key "credentials"
Process rating: all ten parameters 53/100
- 0Result and completion. Does not say what the result is
- 0Inputs and preconditions. Does not say what the process needs to start
- 0Failures and branches. Linear process with no failure handling
- 20When it triggers. No condition that starts the skill
- 30Running it twice. 1 mutating operations with no state check
- 100Tools and files. No external tools needed
- 100Steps. 5 steps
- 100Consistency. Name and required fields are in place
- 100Execution cost. Instruction body is 1071 tokens
- 100Progress reporting. Reports progress
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
- +4Description does not say when NOT to use the skill (false activations)
- +3Output format is not stated: the model decides each time
- +1No license
- +2Single-language instructions
- +3Description length 625: enough signal without eating the budget
- +4Structure: 8 headings
- +3Step-by-step instructions: 5 items
- +4Has examples (3 code blocks)
- +4Reference files are cited in the instructions (2 of 5)
- +3All 1 scripts are documented
Quality base 70; lint remarks subtract, signals add up to 100. Result: 89.