v1-io/v1-deep-research
plugins/v1tamins/skills/v1-deep-research/SKILL.md
Conducts multi-source deep research with iterative refinement and structured synthesis. Triggers on "deep research", "research report", "comprehensive analysis", "analyze X", "compare X vs Y", "state of the art", "what does the research say about". NOT for: simple lookups, debugging, questions answerable in 1-2 searches, code review.
$ install --global
skillsauthnpx skillsauth add v1-io/v1tamins v1-deep-researchInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
Security Scan Results
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
3
Scanners Passed
9
Scanners in report
Clean
TrivyContainer and dependency vulnerability scanner
95%
Clean
SemgrepStatic code analysis for vulnerabilities
95%
Clean
mcp-scan (Snyk)Model Context Protocol security validation
95%
Skipped
Snyk (dep)Open source security scanning
50%
Skipped
Socket.devSupply chain security analysis
50%
Skipped
VirusTotalMulti-engine malware detection
50%
Skipped
CrowdStrikeAdvanced threat intelligence
50%
Skipped
OSV-ScannerOpen Source Vulnerability database check
50%
Skipped
OWASP Dep-Check
50%
Last scanned: May 21, 2026, 2:13 AM152.5s6 files scanned
SKILL.md
- name:
- v1-deep-research
- description:
- >
- NOT for:
- simple lookups, debugging, questions answerable in 1-2 searches, code review.
Deep Research
Autonomous research skill that plans, executes multi-pass search, and synthesizes findings into structured reports with full source attribution.
Decision Tree
User request
|
├─ Answerable in 1-2 searches? ──> STOP. Use WebSearch directly.
├─ Debugging or code question? ──> STOP. Use Explore agent or Read tool.
├─ "Compare X vs Y" detected? ──> COMPARISON mode (see methodology.md)
└─ Research question ──> Select mode:
├─ User says "quick" or simple topic ──> QUICK
├─ Default ──> STANDARD
└─ User says "deep" or "thorough" ──> DEEP
Modes
| Phase | Quick | Standard | Deep | |-------|-------|----------|------| | 1. PLAN | 2 orientation searches, 3 sub-queries | 2 orientation searches, 5 sub-queries | 3 orientation searches, 7 sub-queries | | 2. RESEARCH | Parallel sub-queries only | + 2 sub-agents + 1 gap-fill round | + 3 sub-agents + 2 gap-fill rounds + source curation | | 3. SYNTHESIZE | Direct synthesis, no critique | Outline refinement + critique | Outline refinement + multi-persona critique + delta-queries | | Target length | 2,000-4,000 words | 4,000-8,000 words | 8,000-15,000 words | | Target sources | 8+ | 15+ | 25+ |
Execution
Step 0: Run date +%Y-%m-%d to get the real current date. Use this date in all searches.
Phase 1 -- PLAN: Read reference/methodology.md and follow the PLAN phase instructions.
Phase 2 -- RESEARCH: Continue following reference/methodology.md RESEARCH phase.
Phase 3 -- SYNTHESIZE: Read reference/report-assembly.md for generation strategy. Read reference/quality-gates.md for validation standards. Follow the SYNTHESIZE phase in reference/methodology.md.
Validation: Resolve the bundled validator, then run it up to three times until the report passes:
REPO_ROOT="$(git rev-parse --show-toplevel 2>/dev/null || pwd)"
for dir in \
"$REPO_ROOT/plugins/v1tamins/skills/v1-deep-research" \
"${CLAUDE_PLUGIN_ROOT:-}" \
"$HOME/.codex/skills/v1-deep-research" \
"$HOME/.claude/skills/v1-deep-research"; do
[ -n "$dir" ] && [ -f "$dir/scripts/validate_report.py" ] && SKILL_ROOT="$dir" && break
done
if [ -z "${SKILL_ROOT:-}" ]; then
echo "ERROR: Could not find scripts/validate_report.py" >&2
exit 1
fi
python3 "$SKILL_ROOT/scripts/validate_report.py" --report [path]
Output Contract
Reports MUST include:
- Executive Summary (200-400 words, standalone, no citations)
- Structured body sections with inline citations
[N] - Methodology note (what was searched, how many sources, date range)
- Limitations / gaps acknowledged
- Bibliography with numbered entries matching inline citations
- All sources hyperlinked
Output location: ~/Documents/Research/[Topic]_[YYYYMMDD]/report.md
Flags
--quick/--deep-- override mode selection--context-- emit compact summary for other skills (no full report)--compare-- force comparison mode for "X vs Y" queries
Related Skills
v1-io/v1-testing-prototypes
tools
VerifiedTrustedCommunity
Use when planning or synthesizing user tests for prototypes, mockups, clickable demos, product concepts, design flows, landing pages, or early product specs. Triggers on "test this prototype", "prototype testing", "user test plan", "validate this product idea", "test with users".
5SKILL.mdUpdated Jun 13, 2026
v1-io/v1-html-it
development
VerifiedTrustedCommunity
Use when creating a polished self-contained HTML page would help the user understand, compare, review, present, tune, or interact with information better than Markdown. Triggers on "make an HTML page", "HTML artifact", "nice HTML", "visual report", "interactive explainer", "one-page dashboard", "shareable page".
5SKILL.mdUpdated Jun 13, 2026
v1-io/v1-canon2skill
data-ai
VerifiedTrustedCommunity
Use when turning a textbook, PDF, blog post, article, paper, course, notes, transcript, or other source material into suggested agent skills or skill improvements. Triggers on "what skills could come from this", "extract skills from", "turn this into skills", "skill ideas from this source".
5SKILL.mdUpdated Jun 13, 2026
v1-io/v1-deep-review
development
VerifiedTrustedCommunity
Run an extremely strict maintainability review for abstraction quality, giant files, and spaghetti-condition growth. Use for large prs, new features/architectures, a deep code quality audit, or especially harsh maintainability review.
5SKILL.mdUpdated May 23, 2026