interview-to-spec — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited interview-to-spec (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Interview the user to draft a spec for the task at hand. Cover goal, behaviors, acceptance criteria, edge cases, and out-of-scope — one branch at a time, resolving each before moving on. For every question, offer your own recommended answer.
Ask one question at a time.
If a question can be answered by exploring the codebase, explore it instead.
If during the interview you sense the scope is actually multiple independent features (separate goals, different code paths, would ship as separate PRs), pause and propose a split. Show 2–5 one-line options and ask the user to pick which one to spec first.
Continue the interview for only that one. The others stay outside the repo — the user manages them wherever they keep their backlog (issue tracker, notes, their head). Do not mention them in the spec. Do not add a "follow-up" or "future work" section. Do not create a roadmap file. The spec describes only the work being executed now.
When the user wants to do another one later, they invoke interview-to-spec again with that idea as the input.
When the interview converges, write the spec to spec/<short-task-slug>.md (create the directory if missing). Use these sections:
Keep the spec focused on the what and why. Implementation details belong in a later step.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.