private-company-research

Deep research framework for pre-IPO / private companies (Ant Group, SpaceX, Stripe, ByteDance...). Six analyst lenses — business model, financial forensics, competitive landscape, risk & governance, tech & IP, alternative-data signals — run in parallel via run_swarm, then cross-validated for signal consistency before any verdict. Built around the core challenge of private-company work: information is scarce, so every data point carries a confidence label (high / medium / low), inference is shown separately from fact, and 'I don't know' is a valid output. Outputs a fair-value range, exit-path analysis, and an information-gap map. Use for any unlisted company where you need to judge what the business is actually worth.

Quality

81%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Quality

Content

77%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The content is highly actionable with concrete sources, named tools, specific valuation parameters, and a clear validated workflow culminating in a structured report. Its weaknesses are token efficiency (redundant framing and a restated Key Principles section) and progressive disclosure — a skill this large keeps all lens and template detail inline instead of offloading to reference files.

Suggestions

Split each of the six lens specifications into one reference file each (e.g. references/lens-business.md) and keep SKILL.md as an overview pointing one level deep, which would lift progressive_disclosure to the top anchor.

Remove the 'Key Principles' section or fold its non-redundant items into the relevant lenses, since most points restate the Execution/Cross-Validation/Data Labeling guidance already given.

Trim the 'AI Research Bias Self-Check' framing and framework-characteristics preamble that overlap with the description to reduce restatement and tighten conciseness.

Dimension	Reasoning	Score
Conciseness	The body is largely high-signal domain methodology rather than concepts Claude already knows, but it carries clear redundancy — the 'Key Principles' section restates lens/integration content, and the 'AI Research Bias Self-Check' and framework framing overlap with the description — so it 'could be tightened' per the score-2 anchor.	2 / 3
Actionability	Provides concrete named tools and sources (Google Patents/CNIPA/USPTO, QuestMobile/Sensor Tower, 天眼查/企查查, SharesPost/EquityZen), specific numeric valuation parameters, explicit tool calls (run_swarm, report_audit, write_file) and exact output paths, matching 'fully executable/specific examples'.	3 / 3
Workflow Clarity	Clear parallel-lens → team-lead integrate → cross-validate → report pipeline with a mandatory 'Cross-Validation' section, an explicit 'report_audit' quality gate, and a 14-step ordered report checklist, matching 'clear sequence with explicit validation steps and checklists'.	3 / 3
Progressive Disclosure	No references/scripts/assets bundle exists and the ~155-line skill is a single monolithic file with all six full lens specs and the report template inline; it is well-sectioned but the per-lens detail and template are content that 'should be separate' in reference files, matching the score-2 anchor rather than a one-level-deep split.	2 / 3
	Total	10 / 12 Passed

Description

85%

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description is specific, complete, and distinctive, with an explicit 'Use for' trigger and concrete deliverables. Its main weakness is trigger-term quality: natural user phrasing is partly crowded out by specialized analyst jargon, so it could be broadened with plainer variations.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — six named analyst lenses, parallel execution via run_swarm, cross-validation, and three explicit output artifacts (fair-value range, exit-path analysis, information-gap map) — matching the 'lists multiple specific concrete actions' anchor.	3 / 3
Completeness	Clearly answers both what (the multi-lens framework and its stated outputs) and when via an explicit 'Use for any unlisted company where you need to judge what the business is actually worth.' trigger clause, matching the top anchor.	3 / 3
Trigger Term Quality	Natural terms like 'pre-IPO', 'private companies', 'unlisted company', and example company names are present, but the term set is dominated by analytic jargon ('financial forensics', 'alternative-data signals', 'information-gap map') and lacks broader plain-user variations, matching 'some relevant keywords but missing common variations'.	2 / 3
Distinctiveness Conflict Risk	The pre-IPO / unlisted-company valuation niche is highly specific and clearly distinguishable from generic research or public-company skills, with distinct triggers unlikely to fire for the wrong skill.	3 / 3
	Total	11 / 12 Passed

Validation

93%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	15 / 16 Passed

Repository: HKUDS/Vibe-Trading
Path: agent/src/skills/private-company-research/SKILL.md
Commit: 8643fcd

Reviewed: 2 days ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.