bmad-plus 0.9.0 → 0.9.1

package/osint-agent-package/install.sh

@@ -1,76 +1,76 @@
-#!/bin/bash
-# OSINT Agent — Installation Script (macOS/Linux)
-# Usage: ./install.sh [project_root]
-
-set -euo pipefail
-
-PROJECT_ROOT="${1:-.}"
-PACKAGE_DIR="$(cd "$(dirname "$0")" && pwd)"
-
-echo "========================================"
-echo " OSINT Agent Installer for BMAD"
-echo "========================================"
-echo ""
-
-# Resolve project root
-PROJECT_ROOT="$(cd "$PROJECT_ROOT" && pwd)"
-echo "Project root: $PROJECT_ROOT"
-
-# Verify BMAD
-BMAD_DIR="$PROJECT_ROOT/_bmad/bmm/agents"
-if [ ! -d "$BMAD_DIR" ]; then
-    echo "ERROR: BMAD not found at $BMAD_DIR"
-    echo "Run: npx bmad-method install"
-    exit 1
-fi
-echo "  ✅ BMAD found"
-
-# Verify Python
-if ! command -v python3 &>/dev/null; then
-    echo "ERROR: Python3 not found. Install Python 3.10+"
-    exit 1
-fi
-echo "  ✅ Python found: $(python3 --version)"
-
-# 1. Copy agent definition
-echo ""
-echo "1. Installing agent definition..."
-cp "$PACKAGE_DIR/agents/osint-investigator.md" "$BMAD_DIR/"
-echo "   → osint-investigator.md"
-
-# 2. Copy skills
-echo ""
-echo "2. Installing skills..."
-SKILLS_DIR="$PROJECT_ROOT/.agents/skills"
-
-mkdir -p "$SKILLS_DIR/bmad-osint-investigator"
-cp "$PACKAGE_DIR/skills/bmad-osint-investigator/SKILL.md" "$SKILLS_DIR/bmad-osint-investigator/"
-echo "   → bmad-osint-investigator/"
-
-mkdir -p "$SKILLS_DIR/bmad-osint-investigate/osint/scripts"
-mkdir -p "$SKILLS_DIR/bmad-osint-investigate/osint/references"
-mkdir -p "$SKILLS_DIR/bmad-osint-investigate/osint/assets"
-cp "$PACKAGE_DIR/skills/bmad-osint-investigate/SKILL.md" "$SKILLS_DIR/bmad-osint-investigate/"
-cp "$PACKAGE_DIR/skills/bmad-osint-investigate/osint/SKILL.md" "$SKILLS_DIR/bmad-osint-investigate/osint/"
-cp "$PACKAGE_DIR/skills/bmad-osint-investigate/osint/scripts/"*.py "$SKILLS_DIR/bmad-osint-investigate/osint/scripts/"
-cp "$PACKAGE_DIR/skills/bmad-osint-investigate/osint/references/"* "$SKILLS_DIR/bmad-osint-investigate/osint/references/"
-cp "$PACKAGE_DIR/skills/bmad-osint-investigate/osint/assets/"* "$SKILLS_DIR/bmad-osint-investigate/osint/assets/"
-echo "   → bmad-osint-investigate/ (full pipeline + scripts)"
-
-# 3. Make scripts executable
-chmod +x "$SKILLS_DIR/bmad-osint-investigate/osint/scripts/"*.py
-
-# 4. Run diagnostic
-echo ""
-echo "3. Running diagnostic..."
-python3 "$SKILLS_DIR/bmad-osint-investigate/osint/scripts/diagnose.py"
-
-# 5. Summary
-echo ""
-echo "========================================"
-echo " ✅ Installation complete!"
-echo "========================================"
-echo ""
-echo "Next steps:"
-echo "  1. Configure API keys - see SETUP_KEYS.md"
-echo "  2. Invoke 'bmad-osint-investigator' skill in your AI agent"
+#!/bin/bash
+# OSINT Agent — Installation Script (macOS/Linux)
+# Usage: ./install.sh [project_root]
+
+set -euo pipefail
+
+PROJECT_ROOT="${1:-.}"
+PACKAGE_DIR="$(cd "$(dirname "$0")" && pwd)"
+
+echo "========================================"
+echo " OSINT Agent Installer for BMAD"
+echo "========================================"
+echo ""
+
+# Resolve project root
+PROJECT_ROOT="$(cd "$PROJECT_ROOT" && pwd)"
+echo "Project root: $PROJECT_ROOT"
+
+# Verify BMAD
+BMAD_DIR="$PROJECT_ROOT/_bmad/bmm/agents"
+if [ ! -d "$BMAD_DIR" ]; then
+    echo "ERROR: BMAD not found at $BMAD_DIR"
+    echo "Run: npx bmad-method install"
+    exit 1
+fi
+echo "  ✅ BMAD found"
+
+# Verify Python
+if ! command -v python3 &>/dev/null; then
+    echo "ERROR: Python3 not found. Install Python 3.10+"
+    exit 1
+fi
+echo "  ✅ Python found: $(python3 --version)"
+
+# 1. Copy agent definition
+echo ""
+echo "1. Installing agent definition..."
+cp "$PACKAGE_DIR/agents/osint-investigator.md" "$BMAD_DIR/"
+echo "   → osint-investigator.md"
+
+# 2. Copy skills
+echo ""
+echo "2. Installing skills..."
+SKILLS_DIR="$PROJECT_ROOT/.agents/skills"
+
+mkdir -p "$SKILLS_DIR/bmad-osint-investigator"
+cp "$PACKAGE_DIR/skills/bmad-osint-investigator/SKILL.md" "$SKILLS_DIR/bmad-osint-investigator/"
+echo "   → bmad-osint-investigator/"
+
+mkdir -p "$SKILLS_DIR/bmad-osint-investigate/osint/scripts"
+mkdir -p "$SKILLS_DIR/bmad-osint-investigate/osint/references"
+mkdir -p "$SKILLS_DIR/bmad-osint-investigate/osint/assets"
+cp "$PACKAGE_DIR/skills/bmad-osint-investigate/SKILL.md" "$SKILLS_DIR/bmad-osint-investigate/"
+cp "$PACKAGE_DIR/skills/bmad-osint-investigate/osint/SKILL.md" "$SKILLS_DIR/bmad-osint-investigate/osint/"
+cp "$PACKAGE_DIR/skills/bmad-osint-investigate/osint/scripts/"*.py "$SKILLS_DIR/bmad-osint-investigate/osint/scripts/"
+cp "$PACKAGE_DIR/skills/bmad-osint-investigate/osint/references/"* "$SKILLS_DIR/bmad-osint-investigate/osint/references/"
+cp "$PACKAGE_DIR/skills/bmad-osint-investigate/osint/assets/"* "$SKILLS_DIR/bmad-osint-investigate/osint/assets/"
+echo "   → bmad-osint-investigate/ (full pipeline + scripts)"
+
+# 3. Make scripts executable
+chmod +x "$SKILLS_DIR/bmad-osint-investigate/osint/scripts/"*.py
+
+# 4. Run diagnostic
+echo ""
+echo "3. Running diagnostic..."
+python3 "$SKILLS_DIR/bmad-osint-investigate/osint/scripts/diagnose.py"
+
+# 5. Summary
+echo ""
+echo "========================================"
+echo " ✅ Installation complete!"
+echo "========================================"
+echo ""
+echo "Next steps:"
+echo "  1. Configure API keys - see SETUP_KEYS.md"
+echo "  2. Invoke 'bmad-osint-investigator' skill in your AI agent"

package/osint-agent-package/skills/bmad-osint-investigate/SKILL.md

@@ -1,147 +1,147 @@
----
-name: bmad-osint-investigate
-description: "Run the full OSINT investigation pipeline on a person or company. From a name or handle to a scored dossier with psychoprofile, career map, and confidence grades. Google dorks, French registries, enrichment databases, 55+ Apify actors, 7 search APIs."
----
-
-# OSINT Investigation Skill
-
-This skill executes a systematic OSINT intelligence gathering pipeline on an individual or company.
-
-## Activation Instructions
-
-<investigation-activation CRITICAL="TRUE">
-
-### Step 1: Load the Full Pipeline
-1. READ the complete OSINT pipeline file at: `{skill-dir}/osint/SKILL.md`
-2. READ the enrichment databases reference: `{skill-dir}/osint/references/enrichment-databases-fr.md`
-3. FOLLOW every phase precisely as documented
-
-### Step 2: Get Target Information
-Ask the user for:
-- **Target name** or handle/URL
-- **Type**: Personne physique or Personne morale
-- **Context** (optional): company, city, industry, any known details
-- **Scope**: Full investigation (Phase 0→6) or specific phase only
-- **Budget limit**: Default ≤$0.50, ask for permission above that
-
-### Step 3: Execute the Pipeline
-Follow ALL phases from the OSINT pipeline:
-
-**Phase 0**: Tooling Self-Check
-- Run `python {skill-dir}/osint/scripts/diagnose.py`
-- Report available tools to user
-
-**Phase 1**: Seed Collection (API search)
-- Run parallel search across all available engines
-- Use `python {skill-dir}/osint/scripts/volley.py search "Name" "context"`
-- Merge results with `python {skill-dir}/osint/scripts/volley.py merge <outdir>`
-
-**Phase 1.3**: Google Dorks (FREE — always do this)
-- Search for documents: `"Name" filetype:pdf`, `filetype:pptx`, `filetype:doc`
-- Search for phone: `"Name" ("06" OR "07" OR "+33 6" OR "+33 7")`
-- Search for CV: `"Name" (CV OR resume OR curriculum)`
-- Search for statuts: `"Name" "demeurant" site:pappers.fr`
-- Search for presentations: `"Name" filetype:pptx OR filetype:ppt`
-- Use the browser tool to run these Google searches directly
-- Refer to `{skill-dir}/osint/references/enrichment-databases-fr.md` for all dork templates
-
-**Phase 1.5**: Internal Intelligence
-- Check Telegram history, email, vault contacts BEFORE going external
-
-**Phase 1.7**: French Business Registries (FREE — always do this for FR targets)
-- **Pappers.fr** → SIREN/SIRET, dirigeants, statuts PDF (contain personal addresses!)
-- **Societe.com** → Company fiches, dirigeants, CA
-- **Datalegal.fr** → ALL companies where person is a dirigeant
-- **BODACC** → Official announcements (creation, modification, radiation)
-- **INPI** → Trademarks and patents filed by the person
-- Scrape these pages with: `python {skill-dir}/osint/scripts/jina.py read <url>`
-
-**Phase 2**: Platform Extraction
-- Extract data from LinkedIn, Instagram, Facebook, TikTok, YouTube
-- Use `python {skill-dir}/osint/scripts/apify.py linkedin <url>` or `instagram <handle>`
-- Use `python {skill-dir}/osint/scripts/apify.py run <actor_id> '<json>'` for 55+ Apify actors
-- Refer to `{skill-dir}/osint/references/platforms.md` for platform-specific guide
-- Refer to `{skill-dir}/osint/references/tools.md` for full actor catalog
-
-**Phase 2.5**: Contact Enrichment (for phone numbers and emails)
-- **RocketReach** (rocketreach.co) — Search for target, check visible data (partial phone/email)
-- **ContactOut** (contactout.com) — Check if profile exists with data preview
-- **Kaspr** (kaspr.io) — Best for French mobile numbers via LinkedIn
-- **Apollo.io** — 50 free credits/month, great for B2B enrichment
-- **Hunter.io** — Email patterns for companies
-- Use browser tool to visit these sites and extract visible information
-- Refer to `{skill-dir}/osint/references/enrichment-databases-fr.md` for full list
-
-**Phase 3**: Cross-Reference & Confidence Scoring
-- Build fact table, cross-check key facts, resolve contradictions
-- Assign confidence grades: A (3+ sources), B (2 sources), C (1 source), D (unverified)
-
-**Phase 4**: Psychoprofile
-- Build MBTI/Big Five analysis from content
-- Refer to `{skill-dir}/osint/references/psychoprofile.md` for methodology
-- Refer to `{skill-dir}/osint/references/content-extraction.md` for YouTube/podcast/blog extraction
-
-**Phase 5**: Completeness Evaluation
-- 9 mandatory checks + Depth Score 1-10
-- Gap analysis and stopping criteria
-
-**Phase 6**: Dossier Output
-- For **personnes physiques**: Save to `{project-root}/recherches/personnes-physiques/<nom-prenom>.md`
-- For **personnes morales**: Save to `{project-root}/recherches/personnes-morales/<nom-societe>.md`
-
-### Step 4: Research Escalation
-Always escalate from cheap to expensive:
-- **Level 0** (FREE): Google dorks, Pappers, Societe.com, BODACC, INPI, Datalegal
-- **Level 0.5** (FREE): RocketReach (5/mois), Kaspr (gratuit), Apollo (50/mois), ContactOut
-- **Level 1** (~$0.01): Perplexity Sonar, Tavily, Exa, Jina search
-- **Level 2** (~$0.01): Jina read, Parallel extract
-- **Level 3** (~$0.01-0.10): Apify scrapers, Bright Data
-- **Level 4** (~$0.05-0.50): Perplexity Deep, Exa Deep, Jina Deep
-
-### Available Scripts (Python — stdlib only, zero dependencies)
-All scripts in `{skill-dir}/osint/scripts/`, run with `python <script>`:
-
-| Script | Usage |
-|--------|-------|
-| `diagnose.py` | Self-check of all available tools and API keys |
-| `perplexity.py` | `search <q>`, `sonar <q>`, `deep <q>` |
-| `tavily.py` | `search <q>`, `extract <url>`, `deep <q>` |
-| `exa.py` | `search <q>`, `company <name>`, `people <name>`, `crawl <url>` |
-| `jina.py` | `read <url>`, `search <query>`, `deepsearch <q>` |
-| `parallel.py` | `search <q>`, `extract <url>`, `task <q>` |
-| `apify.py` | `run <actor> <json>`, `linkedin <url>`, `instagram <handle>`, `results`, `store-search` |
-| `brightdata.py` | `scrape <url>`, `search <q>`, `search-yandex <q>` |
-| `volley.py` | `search "Name" "context"` (parallel) / `merge <outdir>` (dedup) |
-| `mcp-client.py` | Lightweight MCP client for Bright Data |
-
-### References
-| File | Content |
-|------|---------|
-| `enrichment-databases-fr.md` | 🇫🇷 French registries, enrichment tools, Google dork templates |
-| `platforms.md` | Social media extraction guides |
-| `tools.md` | 55+ Apify actor catalog |
-| `psychoprofile.md` | MBTI/Big Five methodology |
-| `content-extraction.md` | YouTube/podcast/blog extraction |
-
-### Environment Variables Required
-```
-JINA_API_KEY        — Jina reader/search (recommended first)
-APIFY_API_TOKEN     — Apify scraping (55+ actors)
-PERPLEXITY_API_KEY  — Perplexity Sonar
-TAVILY_API_KEY      — Tavily search
-EXA_API_KEY         — Exa AI
-PARALLEL_API_KEY    — Parallel AI
-BRIGHTDATA_MCP_URL  — Bright Data MCP
-```
-
-</investigation-activation>
-
-## Rules
-- ALWAYS run diagnose.py first to know what tools are available
-- ALWAYS run Google dorks and French registries (Pappers, Societe.com) — they are FREE
-- ALWAYS check enrichment databases (RocketReach, ContactOut) for phone numbers
-- Spend ≤$0.50 without asking — ASK user permission above that
-- Assign confidence grades to EVERY fact (A/B/C/D)
-- Never fabricate data — only report verified findings
-- Escalate from free→paid, fast→deep
-- Save dossiers to `{project-root}/recherches/personnes-physiques/` or `personnes-morales/`
+---
+name: bmad-osint-investigate
+description: "Run the full OSINT investigation pipeline on a person or company. From a name or handle to a scored dossier with psychoprofile, career map, and confidence grades. Google dorks, French registries, enrichment databases, 55+ Apify actors, 7 search APIs."
+---
+
+# OSINT Investigation Skill
+
+This skill executes a systematic OSINT intelligence gathering pipeline on an individual or company.
+
+## Activation Instructions
+
+<investigation-activation CRITICAL="TRUE">
+
+### Step 1: Load the Full Pipeline
+1. READ the complete OSINT pipeline file at: `{skill-dir}/osint/SKILL.md`
+2. READ the enrichment databases reference: `{skill-dir}/osint/references/enrichment-databases-fr.md`
+3. FOLLOW every phase precisely as documented
+
+### Step 2: Get Target Information
+Ask the user for:
+- **Target name** or handle/URL
+- **Type**: Personne physique or Personne morale
+- **Context** (optional): company, city, industry, any known details
+- **Scope**: Full investigation (Phase 0→6) or specific phase only
+- **Budget limit**: Default ≤$0.50, ask for permission above that
+
+### Step 3: Execute the Pipeline
+Follow ALL phases from the OSINT pipeline:
+
+**Phase 0**: Tooling Self-Check
+- Run `python {skill-dir}/osint/scripts/diagnose.py`
+- Report available tools to user
+
+**Phase 1**: Seed Collection (API search)
+- Run parallel search across all available engines
+- Use `python {skill-dir}/osint/scripts/volley.py search "Name" "context"`
+- Merge results with `python {skill-dir}/osint/scripts/volley.py merge <outdir>`
+
+**Phase 1.3**: Google Dorks (FREE — always do this)
+- Search for documents: `"Name" filetype:pdf`, `filetype:pptx`, `filetype:doc`
+- Search for phone: `"Name" ("06" OR "07" OR "+33 6" OR "+33 7")`
+- Search for CV: `"Name" (CV OR resume OR curriculum)`
+- Search for statuts: `"Name" "demeurant" site:pappers.fr`
+- Search for presentations: `"Name" filetype:pptx OR filetype:ppt`
+- Use the browser tool to run these Google searches directly
+- Refer to `{skill-dir}/osint/references/enrichment-databases-fr.md` for all dork templates
+
+**Phase 1.5**: Internal Intelligence
+- Check Telegram history, email, vault contacts BEFORE going external
+
+**Phase 1.7**: French Business Registries (FREE — always do this for FR targets)
+- **Pappers.fr** → SIREN/SIRET, dirigeants, statuts PDF (contain personal addresses!)
+- **Societe.com** → Company fiches, dirigeants, CA
+- **Datalegal.fr** → ALL companies where person is a dirigeant
+- **BODACC** → Official announcements (creation, modification, radiation)
+- **INPI** → Trademarks and patents filed by the person
+- Scrape these pages with: `python {skill-dir}/osint/scripts/jina.py read <url>`
+
+**Phase 2**: Platform Extraction
+- Extract data from LinkedIn, Instagram, Facebook, TikTok, YouTube
+- Use `python {skill-dir}/osint/scripts/apify.py linkedin <url>` or `instagram <handle>`
+- Use `python {skill-dir}/osint/scripts/apify.py run <actor_id> '<json>'` for 55+ Apify actors
+- Refer to `{skill-dir}/osint/references/platforms.md` for platform-specific guide
+- Refer to `{skill-dir}/osint/references/tools.md` for full actor catalog
+
+**Phase 2.5**: Contact Enrichment (for phone numbers and emails)
+- **RocketReach** (rocketreach.co) — Search for target, check visible data (partial phone/email)
+- **ContactOut** (contactout.com) — Check if profile exists with data preview
+- **Kaspr** (kaspr.io) — Best for French mobile numbers via LinkedIn
+- **Apollo.io** — 50 free credits/month, great for B2B enrichment
+- **Hunter.io** — Email patterns for companies
+- Use browser tool to visit these sites and extract visible information
+- Refer to `{skill-dir}/osint/references/enrichment-databases-fr.md` for full list
+
+**Phase 3**: Cross-Reference & Confidence Scoring
+- Build fact table, cross-check key facts, resolve contradictions
+- Assign confidence grades: A (3+ sources), B (2 sources), C (1 source), D (unverified)
+
+**Phase 4**: Psychoprofile
+- Build MBTI/Big Five analysis from content
+- Refer to `{skill-dir}/osint/references/psychoprofile.md` for methodology
+- Refer to `{skill-dir}/osint/references/content-extraction.md` for YouTube/podcast/blog extraction
+
+**Phase 5**: Completeness Evaluation
+- 9 mandatory checks + Depth Score 1-10
+- Gap analysis and stopping criteria
+
+**Phase 6**: Dossier Output
+- For **personnes physiques**: Save to `{project-root}/recherches/personnes-physiques/<nom-prenom>.md`
+- For **personnes morales**: Save to `{project-root}/recherches/personnes-morales/<nom-societe>.md`
+
+### Step 4: Research Escalation
+Always escalate from cheap to expensive:
+- **Level 0** (FREE): Google dorks, Pappers, Societe.com, BODACC, INPI, Datalegal
+- **Level 0.5** (FREE): RocketReach (5/mois), Kaspr (gratuit), Apollo (50/mois), ContactOut
+- **Level 1** (~$0.01): Perplexity Sonar, Tavily, Exa, Jina search
+- **Level 2** (~$0.01): Jina read, Parallel extract
+- **Level 3** (~$0.01-0.10): Apify scrapers, Bright Data
+- **Level 4** (~$0.05-0.50): Perplexity Deep, Exa Deep, Jina Deep
+
+### Available Scripts (Python — stdlib only, zero dependencies)
+All scripts in `{skill-dir}/osint/scripts/`, run with `python <script>`:
+
+| Script | Usage |
+|--------|-------|
+| `diagnose.py` | Self-check of all available tools and API keys |
+| `perplexity.py` | `search <q>`, `sonar <q>`, `deep <q>` |
+| `tavily.py` | `search <q>`, `extract <url>`, `deep <q>` |
+| `exa.py` | `search <q>`, `company <name>`, `people <name>`, `crawl <url>` |
+| `jina.py` | `read <url>`, `search <query>`, `deepsearch <q>` |
+| `parallel.py` | `search <q>`, `extract <url>`, `task <q>` |
+| `apify.py` | `run <actor> <json>`, `linkedin <url>`, `instagram <handle>`, `results`, `store-search` |
+| `brightdata.py` | `scrape <url>`, `search <q>`, `search-yandex <q>` |
+| `volley.py` | `search "Name" "context"` (parallel) / `merge <outdir>` (dedup) |
+| `mcp-client.py` | Lightweight MCP client for Bright Data |
+
+### References
+| File | Content |
+|------|---------|
+| `enrichment-databases-fr.md` | 🇫🇷 French registries, enrichment tools, Google dork templates |
+| `platforms.md` | Social media extraction guides |
+| `tools.md` | 55+ Apify actor catalog |
+| `psychoprofile.md` | MBTI/Big Five methodology |
+| `content-extraction.md` | YouTube/podcast/blog extraction |
+
+### Environment Variables Required
+```
+JINA_API_KEY        — Jina reader/search (recommended first)
+APIFY_API_TOKEN     — Apify scraping (55+ actors)
+PERPLEXITY_API_KEY  — Perplexity Sonar
+TAVILY_API_KEY      — Tavily search
+EXA_API_KEY         — Exa AI
+PARALLEL_API_KEY    — Parallel AI
+BRIGHTDATA_MCP_URL  — Bright Data MCP
+```
+
+</investigation-activation>
+
+## Rules
+- ALWAYS run diagnose.py first to know what tools are available
+- ALWAYS run Google dorks and French registries (Pappers, Societe.com) — they are FREE
+- ALWAYS check enrichment databases (RocketReach, ContactOut) for phone numbers
+- Spend ≤$0.50 without asking — ASK user permission above that
+- Assign confidence grades to EVERY fact (A/B/C/D)
+- Never fabricate data — only report verified findings
+- Escalate from free→paid, fast→deep
+- Save dossiers to `{project-root}/recherches/personnes-physiques/` or `personnes-morales/`