Diagnose and fix a failing or flaky Playwright test. Use whenever a test fails, passes only sometimes, breaks intermittently or needs fixing.

Review what the agent has learned in auto-memory: what to promote to permanent rules, stale entries to prune, duplicates to merge. Use whenever the user runs /si:memory-review or asks what you learned or what to keep.

Make a draft sound less like AI: fix generic, robotic text with rhythm, specifics, personality and brand voice. Use whenever the user says copy sounds like ChatGPT, bland or stiff, or not like them.

Use when designing a cancel flow, save offers, exit surveys or dunning emails, or cutting failed-payment churn. Sizes the revenue at stake with churn_impact_calculator.py first. Health scores and expansion: customer-success-manager.

Use when asked for a landing, marketing or campaign page as a Next.js TSX component with Tailwind. Scaffolds it with landing_page_scaffolder.py using PAS, AIDA or BAB copy, SEO meta and Core Web Vitals targets.

Use when planning demand generation, paid acquisition across Google, LinkedIn and Meta, or a marketing budget, before allocating spend. Computes blended and per-channel CAC with calculate_cac.py and sets MQL to SQL handoffs.

Deep competitor teardown: pricing, reviews, hiring, SEO, UX, scored across 12 dimensions with SWOT and positioning maps. Use whenever the user asks how they compare to a rival or how to respond to a competitor's move.

Use for QMS document control: numbering, versioning, change control or 21 CFR Part 11 e-signatures. Validates documents with document_validator.py and tracks versions with document_version_control.py.

Use for EU MDR 2017/745 work: device classification under Annex VIII, technical file gaps (Annex II/III), clinical evaluation, PMS and PSUR cadence, notified body prep. Checks the file with mdr_gap_analyzer.py.

Use when someone learning Swedish wants a learning path or YouTube clips and podcasts matched to their CEFR level and skill (SFI, Lätt Svenska, Radio Sweden på lätt svenska).

Use when setting up HashiCorp Vault, AWS Secrets Manager, Azure Key Vault or GCP Secret Manager, rotating secrets or auditing secret access. Generates config with vault_config_generator.py, plans rotation with rotation_planner.py and analyzes audit logs.

Edit, proofread and tighten existing marketing copy in focused passes: clarity, voice, proofreading, specificity, readability and AI tells. Use whenever the user asks to polish, clean up, shorten or fix the wording of a page, email or ad.

Use when someone shares SaaS revenue or customer numbers, or asks about ARR, MRR, churn, LTV, CAC, NRR or how their SaaS is doing. Computes metrics with metrics_calculator.py, quick ratio and unit economics.

Analyze, research or value a listed stock or company (US, global, NSE or BSE) with verified data, ratios, scores and valuation. Use whenever asked if a stock is a buy, what it is worth, or how a company looks.

Use for any Terraform module, state backend, provider setup or IaC security review, before editing .tf files. Analyzes modules with tf_module_analyzer.py and scans for misconfigurations with tf_security_scanner.py.

Use for ISO/IEC 42001 AI management system work: gap analysis against Clauses 4-10, the AI risk register and Annex A controls, or the internal audit plan. Runs aims_gap_analyzer.py, ai_risk_register_builder.py and aims_audit_scheduler.py.

Use when planning, promoting or fixing a webinar or virtual event (registrations, show-up, follow-up, conversion). Scores each funnel stage with webinar_funnel_scorer.py to find where it leaks.

Use when planning or running an authorized red team engagement or attack path analysis. Plans the kill chain with engagement_planner.py: MITRE ATT&CK techniques, choke points, OPSEC risk and crown jewels.

Use when hunting threats, sweeping IOCs or looking for anomalies in logs and telemetry. Ranks ATT&CK-mapped signals and z-score anomalies with threat_signal_analyzer.py.

Reply like a real person, direct and warm, without lists or hedging. Use whenever the user says you sound robotic, asks you to be real or talk like a person, wants mirror mode, or is going through grief, job loss or fear. Not for code or factual lookups.