aiskillstore/marketplace / skills/199-biotechnologies/deep-research
deep-research
VerifiedUse when the user needs multi-source research with citation tracking, evidence persistence, and structured report generation. Triggers on "deep research", "comprehensive analysis", "research report", "compare X vs Y", "analyze trends", or "state of the art". Not for simple lookups, debugging, or questions answerable with 1-2 searches.
| File | Kind | Size |
|---|---|---|
| .gitignore | other | 264 B |
| README.md | markdown | 4.8 KB |
| SKILL.md | skill md | 4.2 KB |
| requirements.txt | manifest | 579 B |
| skill-report.json | config | 23.6 KB |
skill-report.json
View on GitHub{
"schema_version": "2.0",
"meta": {
"generated_at": "2026-07-04T14:53:42.444Z",
"slug": "199-biotechnologies-deep-research",
"source_url": "https://github.com/199-biotechnologies/claude-deep-research-skill/tree/main/",
"source_ref": "main",
"model": "codex",
"analysis_version": "3.0.0",
"source_type": "community",
"content_hash": "0c8e777eb13d37b4d5130f59c9237a82528224ba31b58b774cd490a742df8620",
"tree_hash": "f32d8b813a1db20c83751375acf47d5cbf7720bae8633038e608f4bf5c899f13",
"provenance": {
"model_effective": "codex",
"fallback_chain": [
{
"agent": "codex",
"outcome": "success"
}
],
"analysis_version": "3.0.0"
}
},
"skill": {
"name": "deep-research",
"description": "Guides Claude, Codex, and Claude Code through structured multi-source research with citation tracking, evidence records, claim verification, and report packaging.",
"summary": "Create citation-tracked research reports from multi-source evidence and explicit quality gates.",
"icon": "π¦",
"version": "1.0.1",
"author": "199-biotechnologies",
"license": "MIT",
"category": "research",
"tags": [
"deep research",
"citations",
"evidence tracking",
"research reports",
"source verification"
],
"supported_tools": [
"claude",
"codex",
"claude-code"
],
"risk_factors": [
"network",
"filesystem",
"external_commands"
]
},
"security_audit": {
"risk_level": "medium",
"is_blocked": false,
"safe_to_publish": true,
"analysis_status": "ok",
"summary": "Most Markdown backtick and install-path findings are false positives. Three script command references are confirmed because they direct agents to run relative Python scripts with user-supplied paths. No prompt injection evidence found, but provider configuration may be persisted in report artifacts.",
"remediation": [
{
"issue": "Relative Python script commands in the skill instructions",
"suggestion": "Use bundled script paths or require explicit user confirmation before running validation commands. Validate report paths before execution.",
"severity": "medium"
},
{
"issue": "Provider configuration can be written to report artifacts",
"suggestion": "Redact API keys and tokens from run_manifest.json. Store only provider names and nonsecret settings.",
"severity": "medium"
}
],
"risk_factor_evidence": [
{
"factor": "network",
"evidence": [
{
"file": "README.md",
"line_start": 9,
"line_end": 9
}
]
},
{
"factor": "filesystem",
"evidence": [
{
"file": "README.md",
"line_start": 9,
"line_end": 9
},
{
"file": "README.md",
"line_start": 9,
"line_end": 9
}
]
},
{
"factor": "external_commands",
"evidence": [
{
"file": "SKILL.md",
"line_start": 18,
"line_end": 29
},
{
"file": "SKILL.md",
"line_start": 29,
"line_end": 68
},
{
"file": "SKILL.md",
"line_start": 68,
"line_end": 69
},
{
"file": "SKILL.md",
"line_start": 69,
"line_end": 70
},
{
"file": "SKILL.md",
"line_start": 70,
"line_end": 86
},
{
"file": "SKILL.md",
"line_start": 86,
"line_end": 88
},
{
"file": "SKILL.md",
"line_start": 88,
"line_end": 89
},
{
"file": "SKILL.md",
"line_start": 89,
"line_end": 90
},
{
"file": "SKILL.md",
"line_start": 90,
"line_end": 91
},
{
"file": "SKILL.md",
"line_start": 91,
"line_end": 97
}
]
}
],
"static_findings": [
{
"id": "sensitive:.gitignore:22:sqlite-database-file",
"category": "sensitive",
"severity": "medium",
"pattern": "SQLite database file",
"file": ".gitignore",
"line_start": 22,
"line_end": 22,
"snippet": "Thumbs.db"
},
{
"id": "network:README.md:9:hardcoded-url",
"category": "network",
"severity": "low",
"pattern": "Hardcoded URL",
"file": "README.md",
"line_start": 9,
"line_end": 9,
"snippet": "git clone https://github.com/199-biotechnologies/claude-deep-research-skill.git ~/.claude/skills/dee"
},
{
"id": "filesystem:README.md:9:hidden-file-in-home-directory",
"category": "filesystem",
"severity": "high",
"pattern": "Hidden file in home directory",
"file": "README.md",
"line_start": 9,
"line_end": 9,
"snippet": "git clone https://github.com/199-biotechnologies/claude-deep-research-skill.git ~/.claude/skills/dee"
},
{
"id": "filesystem:README.md:9:hidden-file-access",
"category": "filesystem",
"severity": "medium",
"pattern": "Hidden file access",
"file": "README.md",
"line_start": 9,
"line_end": 9,
"snippet": "git clone https://github.com/199-biotechnologies/claude-deep-research-skill.git ~/.claude/skills/dee"
},
{
"id": "external_commands:SKILL.md:18:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 18,
"line_end": 29,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:29:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 29,
"line_end": 68,
"snippet": "```"
},
{
"id": "external_commands:SKILL.md:68:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 68,
"line_end": 69,
"snippet": "- `python scripts/validate_report.py --report [path]`"
},
{
"id": "external_commands:SKILL.md:69:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 69,
"line_end": 70,
"snippet": "- `python scripts/verify_citations.py --report [path]`"
},
{
"id": "external_commands:SKILL.md:70:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 70,
"line_end": 86,
"snippet": "- `python scripts/md_to_html.py [markdown_path]`"
},
{
"id": "external_commands:SKILL.md:86:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 86,
"line_end": 88,
"snippet": "**Output files (all to `~/Documents/[Topic]_Research_[YYYYMMDD]/`):**"
},
{
"id": "external_commands:SKILL.md:88:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 88,
"line_end": 89,
"snippet": "- `sources.jsonl` β stable source registry with canonical IDs"
},
{
"id": "external_commands:SKILL.md:89:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 89,
"line_end": 90,
"snippet": "- `evidence.jsonl` β append-only evidence store with quotes and locators"
},
{
"id": "external_commands:SKILL.md:90:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 90,
"line_end": 91,
"snippet": "- `claims.jsonl` β atomic claim ledger with support status"
},
{
"id": "external_commands:SKILL.md:91:ruby-shell-backtick-execution",
"category": "external_commands",
"severity": "medium",
"pattern": "Ruby/shell backtick execution",
"file": "SKILL.md",
"line_start": 91,
"line_end": 97,
"snippet": "- `run_manifest.json` β query, mode, assumptions, provider config"
}
],
"finding_verdicts": [
{
"id": "sensitive:.gitignore:22:sqlite-database-file",
"verdict": "false_positive",
"confidence": 0.98,
"reason": ".gitignore line 22 contains Thumbs.db, an operating system cache file name. It is not a SQLite database file or sensitive data."
},
{
"id": "network:README.md:9:hardcoded-url",
"verdict": "false_positive",
"confidence": 0.93,
"reason": "README.md line 9 shows a GitHub clone URL for installation. It is documentation for fetching the skill, not a runtime network request or exfiltration path."
},
{
"id": "filesystem:README.md:9:hidden-file-in-home-directory",
"verdict": "false_positive",
"confidence": 0.88,
"reason": "README.md line 9 installs the skill into the standard Claude skills directory under ~/.claude. This is an expected manual installation path, not hidden data access."
},
{
"id": "filesystem:README.md:9:hidden-file-access",
"verdict": "false_positive",
"confidence": 0.88,
"reason": "The hidden path is the documented Claude skills installation directory. The line does not instruct reading, modifying, or exfiltrating unrelated hidden files."
},
{
"id": "external_commands:SKILL.md:18:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.97,
"reason": "SKILL.md line 18 opens a Markdown code fence for a decision tree. It is not Ruby backtick execution and contains no shell command."
},
{
"id": "external_commands:SKILL.md:29:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.97,
"reason": "SKILL.md line 29 closes a Markdown code fence. The surrounding block is static workflow text, not executable Ruby or shell code."
},
{
"id": "external_commands:SKILL.md:68:ruby-shell-backtick-execution",
"verdict": "confirmed",
"confidence": 0.66,
"reason": "SKILL.md line 68 lists a Python validation command with a user-supplied report path. Relative script execution can run local code and process sensitive report content.",
"severity": "medium"
},
{
"id": "external_commands:SKILL.md:69:ruby-shell-backtick-execution",
"verdict": "confirmed",
"confidence": 0.66,
"reason": "SKILL.md line 69 lists a Python citation verification command with a user-supplied report path. This is legitimate tooling, but it is still local command execution.",
"severity": "medium"
},
{
"id": "external_commands:SKILL.md:70:ruby-shell-backtick-execution",
"verdict": "confirmed",
"confidence": 0.66,
"reason": "SKILL.md line 70 lists a Python HTML conversion command with a user-supplied Markdown path. The relative script path and document input create command execution risk.",
"severity": "medium"
},
{
"id": "external_commands:SKILL.md:86:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.92,
"reason": "SKILL.md line 86 documents an output folder path in inline Markdown. It is filesystem output guidance, not Ruby or shell backtick execution."
},
{
"id": "external_commands:SKILL.md:88:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.95,
"reason": "SKILL.md line 88 names sources.jsonl as an output artifact. It is a file name in documentation, not an executable command."
},
{
"id": "external_commands:SKILL.md:89:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.95,
"reason": "SKILL.md line 89 names evidence.jsonl as an output artifact. It is a documented data file, not command execution."
},
{
"id": "external_commands:SKILL.md:90:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.95,
"reason": "SKILL.md line 90 names claims.jsonl as an output artifact. It is not a shell command or Ruby backtick expression."
},
{
"id": "external_commands:SKILL.md:91:ruby-shell-backtick-execution",
"verdict": "false_positive",
"confidence": 0.94,
"reason": "SKILL.md line 91 names run_manifest.json and describes its contents. The backticks mark a file name, not executable code."
}
],
"semantic_findings": [
{
"title": "Potential Secret Persistence in Research Manifest",
"description": "The README tells users to configure search provider keys, and the output contract stores provider config in run_manifest.json. Provider secrets should not be written to report artifacts.",
"severity": "medium",
"locations": [
{
"file": "README.md",
"line_start": 16,
"line_end": 20
},
{
"file": "SKILL.md",
"line_start": 86,
"line_end": 91
}
],
"confidence": 0.67,
"confidence_reasoning": "README.md line 20 explicitly references provider keys, and SKILL.md line 91 says provider config is stored. The text does not prove keys are persisted, so confidence is medium."
}
],
"critical_findings": [],
"high_findings": [],
"medium_findings": [
{
"title": "Ruby/shell backtick execution",
"description": "- `python scripts/validate_report.py --report [path]`",
"locations": [
{
"file": "SKILL.md",
"line_start": 68,
"line_end": 69
}
],
"review_kind": "capability",
"source_category": "external_commands",
"source_severity": "medium",
"confidence": 0.66,
"confidence_reasoning": "SKILL.md line 68 lists a Python validation command with a user-supplied report path. Relative script execution can run local code and process sensitive report content."
},
{
"title": "Ruby/shell backtick execution",
"description": "- `python scripts/verify_citations.py --report [path]`",
"locations": [
{
"file": "SKILL.md",
"line_start": 69,
"line_end": 70
}
],
"review_kind": "capability",
"source_category": "external_commands",
"source_severity": "medium",
"confidence": 0.66,
"confidence_reasoning": "SKILL.md line 69 lists a Python citation verification command with a user-supplied report path. This is legitimate tooling, but it is still local command execution."
},
{
"title": "Ruby/shell backtick execution",
"description": "- `python scripts/md_to_html.py [markdown_path]`",
"locations": [
{
"file": "SKILL.md",
"line_start": 70,
"line_end": 86
}
],
"review_kind": "capability",
"source_category": "external_commands",
"source_severity": "medium",
"confidence": 0.66,
"confidence_reasoning": "SKILL.md line 70 lists a Python HTML conversion command with a user-supplied Markdown path. The relative script path and document input create command execution risk."
},
{
"title": "Potential Secret Persistence in Research Manifest",
"description": "The README tells users to configure search provider keys, and the output contract stores provider config in run_manifest.json. Provider secrets should not be written to report artifacts.",
"locations": [
{
"file": "README.md",
"line_start": 16,
"line_end": 20
},
{
"file": "SKILL.md",
"line_start": 86,
"line_end": 91
}
],
"review_kind": "security",
"source_category": "semantic",
"source_severity": "medium",
"confidence": 0.67,
"confidence_reasoning": "README.md line 20 explicitly references provider keys, and SKILL.md line 91 says provider config is stored. The text does not prove keys are persisted, so confidence is medium."
}
],
"low_findings": [],
"dangerous_patterns": [],
"files_scanned": 4,
"total_lines": 276,
"audit_model": "codex",
"audited_at": "2026-07-04T14:53:42.444Z",
"agent_auto_install_policy": "allowed",
"manual_install_policy": "allowed"
},
"content": {
"user_title": "Build Citation-Tracked Research Reports",
"value_statement": "Complex research can lose source context, assumptions, and claim support. This skill guides Claude, Codex, and Claude Code through structured evidence collection and report assembly.",
"seo_keywords": [
"Claude",
"Codex",
"Claude Code",
"deep research skill",
"citation tracking",
"research report generator",
"source verification",
"evidence ledger",
"multi-source analysis",
"claim validation"
],
"actual_capabilities": [
"Guides quick, standard, deep, and ultradeep research modes.",
"Defines a phased workflow from scope and planning through retrieval, synthesis, critique, and packaging.",
"Requires cited findings, a complete bibliography, and explicit limitations.",
"Specifies evidence files for sources, quotes, claims, and run metadata.",
"Lists optional validation commands for report structure, citations, and HTML conversion."
],
"limitations": [
"Best suited for complex research, not simple lookups or debugging.",
"Requires available search tools or web access for current source collection.",
"Referenced helper scripts and templates may need separate packaging or setup.",
"Outputs may contain sensitive research context if users provide confidential topics."
],
"use_cases": [
{
"title": "Technology Evaluation",
"description": "Compare platforms, methods, or vendors with cited evidence and decision-focused recommendations.",
"target_user": "Engineering leader"
},
{
"title": "Scientific Literature Review",
"description": "Map recent findings, source quality, and unresolved questions for a technical topic.",
"target_user": "Research analyst"
},
{
"title": "Market Landscape Report",
"description": "Synthesize trends, competitors, adoption signals, and risks into a structured report.",
"target_user": "Product strategist"
}
],
"prompt_templates": [
{
"title": "Research a Topic",
"prompt": "Use deep research mode to explain [topic] for [audience]. Include current sources, key findings, limitations, and a bibliography.",
"scenario": "Beginner request for a structured overview."
},
{
"title": "Compare Two Options",
"prompt": "Compare [option A] and [option B] for [decision context]. Use cited evidence, tradeoffs, risks, and a recommendation.",
"scenario": "Intermediate request for a decision report."
},
{
"title": "Analyze Recent Trends",
"prompt": "Research trends in [field] from the last [time period]. Separate confirmed evidence, emerging signals, and uncertain claims.",
"scenario": "Advanced request for time-sensitive analysis."
},
{
"title": "Run an Ultradeep Review",
"prompt": "Run ultradeep research on [complex question]. Build an evidence ledger, challenge assumptions, verify major claims, and produce a final report.",
"scenario": "Advanced request for rigorous synthesis and critique."
}
],
"output_examples": [
{
"input": "Deep research the current state of quantum computing for enterprise planning.",
"output": "A cited report with an executive summary, major findings, limitations, recommendations, and a full bibliography."
},
{
"input": "Compare PostgreSQL and Supabase for our application stack.",
"output": [
"Decision-focused comparison across architecture, operations, security, cost, and team fit.",
"Clear recommendation with assumptions and cited support."
]
},
{
"input": "Analyze recent biotech AI platform trends for a product strategy review.",
"output": "A market landscape report that separates mature evidence, emerging signals, and open questions."
}
],
"best_practices": [
"State the audience, decision context, and time horizon before starting.",
"Ask for citations and limitations to remain visible in the final report.",
"Review generated evidence files for confidential data before sharing outputs."
],
"anti_patterns": [
"Using the skill for one-answer lookups that need only a quick search.",
"Accepting uncited claims in the final report.",
"Storing provider keys or confidential source material in report artifacts."
],
"faq": [
{
"question": "Does this replace a search engine?",
"answer": "No. It organizes multi-source research and uses available search tools to gather evidence."
},
{
"question": "Can it produce HTML or PDF?",
"answer": "The skill describes HTML and PDF packaging, but local tool setup may be required."
},
{
"question": "Does it need API keys?",
"answer": "Basic use may work without extra keys. Multi-provider search needs configured provider keys."
},
{
"question": "Can it verify citations?",
"answer": "It defines citation checks and optional validation commands. Users should still review important citations."
},
{
"question": "Where are outputs stored?",
"answer": "The skill directs outputs to a research folder under Documents and uses evidence ledgers."
},
{
"question": "Is it suitable for confidential research?",
"answer": "Use caution. Review generated files and redact secrets before sharing or syncing outputs."
}
]
},
"file_structure": [
{
"name": ".gitignore",
"type": "file",
"path": ".gitignore",
"lines": 32
},
{
"name": "README.md",
"type": "file",
"path": "README.md",
"lines": 120
},
{
"name": "requirements.txt",
"type": "file",
"path": "requirements.txt",
"lines": 15
},
{
"name": "SKILL.md",
"type": "file",
"path": "SKILL.md",
"lines": 109
}
]
}
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| add-product-e2e-evalpaperclipai/paperclipAdd or extend a Paperclip full-stack runner E2E workflow, fixture, matcher, or report evidence path for local or Daytona execution. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| add-runner-evalpaperclipai/paperclipAdd or extend a Paperclip Runner protocol evaluation definition, roster, assertion, or report fixture with provenance and narrow validation. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| diagnose-why-work-stoppedpaperclipai/paperclipCompatibility entry point for failed, stalled, looping, or over-recovered Paperclip work. Use fix-failed-paperclip-run for the full diagnosis, regression, repair, real verification, and green-PR workflow; honor an explicitly requested diagnosis-only scope. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| fix-failed-paperclip-runpaperclipai/paperclipDiagnose and fix failed, stuck, slow, or incorrectly blocked Paperclip tasks and runs. Use with a task/run URL or failure report to trace the root cause, prove a regression red then green, check second-order effects, verify the real workflow, and deliver a PR with passing checks. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |