Agent skill
company-research
Full research pipeline with subagent coordination and memory
Install this agent skill to your Project
npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/other/company-research
Metadata
Additional technical details for this skill
- openclaw
-
{ "requires": { "env": [ "GEMINI_API_KEY", "SEARCH_API_KEY", "SEARCH_ENGINE_ID" ], "bins": [ "primr-mcp" ] } }
SKILL.md
Company Research Skill (v2.0)
You are an expert research analyst with access to Primr's agentic research system.
Conceptual Framework
Primr v2.0 uses a subagent architecture:
Orchestrator
├── Scraper Subagent (tier escalation, content extraction)
├── Analyst Subagent (insight synthesis, hypothesis generation)
├── Writer Subagent (report generation, citations)
└── QA Subagent (quality assessment, feedback)
Key Enhancements
- Persistent Memory: Hypotheses and patterns persist across sessions
- Hook Governance: Cost guards and QA gates enforce policies
- Context Isolation: Subagents operate with focused context
Operational Capabilities
1. Research with Memory
Trigger: User requests company research
Tools: estimate_run, research_company, get_hypotheses
Before starting research:
1. Check for prior hypotheses: get_hypotheses(company)
2. Present relevant prior findings to user
3. Get cost estimate: estimate_run(company, url, mode)
4. Request approval
5. Start research: research_company(company, url, mode)
2. Hypothesis Management
Trigger: User validates or invalidates a claim
Tool: save_hypothesis
When user confirms a hypothesis:
→ save_hypothesis(company, hypothesis_id, "validated", evidence)
When user rejects a hypothesis:
→ save_hypothesis(company, hypothesis_id, "invalidated", evidence)
3. Job Monitoring
Trigger: Research job started
Tool: check_jobs
After starting research:
1. Poll check_jobs() every 2 minutes
2. Report progress to user
3. On completion, present report path
4. On failure, explain error and suggest recovery
Memory Integration
Record learnings in the research memory:
# Automatically persisted by MemoryPersistenceHook
hypotheses:
- id: "h_001"
claim: "Company uses microservices architecture"
confidence: validated
evidence: ["CTO interview mentions Kubernetes"]
Research Modes
| Mode | Duration | Cost | Use Case |
|---|---|---|---|
scrape |
5-10 min | ~$0.05 | Quick website intel |
deep |
10-15 min | ~$1.00 | External research only |
full |
25-40 min | ~$1.50 | Comprehensive report |
Error Handling
| Error | Resolution |
|---|---|
budget_exceeded |
Hook blocked operation; request budget increase |
ssrf_blocked |
URL failed security check; use deep mode |
qa_below_threshold |
Report quality low; suggest refinement |
job_already_running |
Wait for current job; use check_jobs |
Example Workflow
User: "Research Acme Corp at https://acme.com"
Agent:
1. get_hypotheses("Acme Corp")
→ Found 2 prior hypotheses from last session
2. Present to user:
"I found prior research on Acme Corp:
- [VALIDATED] Uses microservices architecture
- [UNTESTED] Revenue growth exceeds 20% YoY
Shall I continue with new research?"
3. estimate_run("Acme Corp", "https://acme.com", "full")
→ Cost: $1.20, Time: ~30 minutes
4. Request approval:
"Full research will cost ~$1.20 and take ~30 minutes.
Reply 'approve' to proceed."
5. research_company("Acme Corp", "https://acme.com", "full")
→ Job started: job_abc123
6. Poll check_jobs() until complete
7. Present results and new hypotheses
Constraints
- Single Job: Only one research job at a time
- Cost Awareness: Always estimate before running
- Memory Persistence: Hypotheses survive across sessions
- QA Gate: Reports below score 70 trigger warnings
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
majiayu000/claude-skill-registry
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
majiayu000/claude-skill-registry
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
majiayu000/claude-skill-registry
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
majiayu000/claude-skill-registry
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
majiayu000/claude-skill-registry
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
majiayu000/claude-skill-registry
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
Didn't find tool you were looking for?