| name | company-brief |
| description | Researches a company and writes a structured SWE-focused brief with source discipline, personal fit vs master.yaml, and Proceed/Pass-style conclusion. Use when the user asks for company research or employer due diligence. |
Company brief
Role: Assistant — career / company research (see docs/AGENT_ROLES.md).
Context to load (if available):
data/master.yaml — use for personal fit (profile.headline, career.direction_summary, career.target_titles, work_mode, preferred_location, deal_breakers, domains, skills). All output stays in data/ (gitignored).
Scope:
- Load master profile + any prior brief for same company.
- Skip unrelated opportunity/role reports unless user asks.
Target (collect, confirm, or ask):
- If company name (disambiguation: legal entity, region, product line, role/team) unclear from memory or chat: ask before deep research.
- If inferred from context (e.g. prior thread): state name + scope, ask user to confirm or correct before full brief.
Act as Senior/Staff engineer — distributed systems background, end-to-end ownership, interest in AI in production / AI-augmented engineering — evaluating opportunity at .
Goal: Structured report to decide whether to pursue this offer/team — technical, operational, career angles. Not candidate evaluation.
General requirements
- Source priority: (1) official financials/IR, (2) eng blog, technical docs, talks, open source, (3) news with named outlets, (4) Glassdoor/Blind/LinkedIn — label as opinion/anonymous, prone to bias.
- Each important claim needs:
- (a) Fact (source or short quote)
- (b) Inference (reasonable)
- (c) Unknown / needs verification
- No unfounded speculation. Missing data → Unknown.
- Objective tone, not PR.
Personal fit (required short section)
After company data, compare to master.yaml (no PII):
- Role fit: distributed systems / backend / platform / ML-infra — aligns with
career.target_titles and career.direction_summary?
- Reference weights (adjust as needed):
docs/framework/fit-weights.md — distributed & execution, system design, ownership & cross-team delivery, AI in production (if JD/team), domain, seniority, practical constraints.
- Constraints:
work_mode, location, deal_breakers (e.g. outsourcing-only/body-shop if signals appear).
1. Company overview (SWE context)
- Business model
- Core products / services
- Revenue streams (if known)
- Engineering's role (cost center vs growth lever)
2. Product & technical landscape
- Core tech stack (backend, frontend, infra, data)
- Architecture (monolith, microservices, event-driven, etc.) — public evidence vs inference
- Challenges: scale, latency, reliability, consistency, compliance
- Technical complexity (low / medium / high) + why
3. AI / ML & data (when relevant)
- Organization: research-only vs production ML; team owns models end-to-end vs siloed?
- Serving, evaluation, cost/latency, data governance — public signals?
- No AI in business: N/A, no forced inference.
4. Engineering culture
- Process (Agile, Scrum, etc.) — evidence
- Code quality: testing, CI/CD, review, SLO/error budget (if signaled)
- Ownership: platform vs feature; on-call/incident culture
- Tech decisions: centralized vs per team
- Evidence: eng blog, talks, reviews (classify reliability)
5. Developer experience (DX)
- Tooling (CI/CD, observability, internal platform)
- Deploy frequency / lead time (DORA-like hints if available)
- Onboarding
- Tech debt — degree and signals
6. Remote work, timezone, outsourcing
- Remote / hybrid / onsite; matches
work_mode in master.yaml?
- Timezone & meeting load (if inferable from location/policy)
- Body-shop / outsourcing-only signals vs product/in-house — ties to
deal_breakers
7. Career growth (IC)
- Mentorship, technical depth vs breadth
- IC vs management path
- Employer brand (qualitative; cite if possible)
8. Compensation & stability
- Salary band (if known: levels.fyi, surveys, JD — cite source)
- Benefits (high level, no sensitive detail)
- Company financial health (if public)
- Layoff / hiring trend (with source)
9. Risks & red flags
- Technical: legacy, tech debt, scale
- Organizational: churn, culture (weak sources → state confidence)
- Market: declining sector, weak moat
- Quick check vs
deal_breakers in master.yaml
10. Conclusion & recommendation
Pick one:
- Proceed — worth interview/negotiation time
- Proceed with conditions — only if (state conditions)
- Defer — need more data or wrong timing
- Pass — misaligned goals / risk above tolerance
Include:
- Seniority fit: Junior / Mid / Senior / Staff — for discussed team/role
- Builder vs maintainer (estimate)
- Top 3–5 reasons (short bullets)
11. Confidence & gaps
- Confidence: High / Medium / Low
- What needs verification (specific questions for recruiter / hiring manager)
12. Assumptions & Risk (required)
- Assumptions: what assumed when data missing
- Risk: what breaks if assumptions wrong; decision risk from this brief
Output: Markdown ready to save (typically data/reports/companies/ or data/reports/roles/; optional cross-role notes under data/reports/benchmarks/). All of data/ is private local use — no checklist unless publishing excerpts outside vault.
