templates/skills/project/fast/pulse-scan/SKILL.md
Run a timeboxed research scan after fast-init to confirm critical constraints and escalation triggers
npx skillsauth add ozerohax/assistagents project-fast-pulse-scanInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
<when_to_use> <item importance="critical">Immediately after fast-init and before stack/architecture decisions</item> <item importance="high">When integrations, compliance, SLA, or external-system dependencies exist</item> </when_to_use>
<required_preload> <item>web-strategy</item> <item>planning-risk-assessment</item> <item>planning-impact-analysis</item> </required_preload>
<inputs> <required>Goal, scope, and constraints from fast-init</required> <optional>Current technical choices and known project pain points</optional> <optional>External standards/documentation/market constraints</optional> </inputs> <method> <step>Formulate 1-3 verifiable research questions</step> <step>Run rapid research using primary sources, limited to critical topics</step> <step>Identify top risks and top constraints impacting architecture and scope</step> <step>Separate facts from hypotheses and note confidence for key findings</step> <step>Capture escalation triggers from fast to standard</step> </method><output_format>
<section>Research questions</section> <section>Verified findings</section> <section>Top constraints</section> <section>Top risks + mitigations</section> <section>Escalation triggers</section> <section>Sources</section> </output_format><quality_rules> <rule importance="critical">Each key finding is backed by at least one reliable source</rule> <rule importance="high">Facts and assumptions are explicitly separated</rule> <rule importance="high">Risks are prioritized by impact on the fast plan</rule> </quality_rules>
<do_not> <item importance="critical">Do not run broad research without a narrowly scoped question</item> <item importance="high">Do not pull secondary risks into the fast plan</item> </do_not>
testing
Visual regression (baseline diffs, UI stabilization)
testing
Design detailed reproducible test cases (Given/When/Then, one check per case); not smoke/regression checklists
development
Mobile testing (native/hybrid/web, gestures, offline, interruptions)
testing
Load testing (k6/JMeter scenarios, benchmarks)