tech-intel

name	tech-intel
description	Technology intelligence gathering. Web research for dependency updates, security advisories, and ecosystem developments relevant to your stack.
user-invocable	true
disable-model-invocation	true
allowed-tools	Bash, Read, Write, WebSearch, WebFetch, Glob
catalog_description	Dependency updates, security advisories, and ecosystem intelligence.

Tech Intel — Technology Intelligence Digest

Researches technology developments relevant to your project stack. Reads CC_LANGUAGE, CC_DATABASE, and CC_MCP_SERVERS from cognitive-core.conf to focus research on your actual dependencies.

Arguments

• $ARGUMENTS -- optional: --date=YYYY-MM-DD, --topic=<focus>
• Topics: claude, stack, security, all (default: all)

Research Topics

Priority 1: Claude Code and Anthropic (DAILY)

• Claude Code CLI releases and changelog
• New model capabilities
• Skills and agents architecture changes
• MCP server ecosystem updates

Priority 2: Project Stack (WEEKLY)

Based on CC_LANGUAGE and CC_DATABASE:

• Language runtime releases and security patches
• Framework and ORM updates
• Database driver and tooling changes
• Notable new packages in the ecosystem

Priority 3: Industry Trends (WEEKLY)

• AI coding assistant developments
• Agentic development patterns
• Developer productivity research

Instructions

Step 1: Research

Search for recent developments (last 24h for daily, last 7 days for weekly):

• "claude code" changelog OR release OR update
• anthropic claude model update [current year]
• Stack-specific queries derived from CC_LANGUAGE and CC_DATABASE
• Security advisories for project dependencies

Step 2: Analyze and Prioritize

For each finding, assess:

• Relevance: Does it affect this project? (HIGH/MEDIUM/LOW)
• Urgency: Act now? (IMMEDIATE/SOON/INFORMATIONAL)
• Impact: What changes if adopted? (ARCHITECTURE/WORKFLOW/COSMETIC)

Step 3: Generate Digest

Save to ${CC_SESSION_DOCS_DIR:-docs}/tech-intel/YYYY-MM-DD.md:

# Tech Intel Digest - YYYY-MM-DD

## Headlines
- [Most important finding]
- [Second finding]

## Claude Code and Anthropic
### [Finding Title]
**Relevance**: HIGH | **Urgency**: IMMEDIATE
**Source**: [URL]
**Summary**: [2-3 sentences]
**Action**: [What to do, if anything]

## Project Stack
### [Finding Title]
...

## Security Advisories
### [Finding Title]
...

## Recommended Reading
1. [Title](URL) - [Why it matters] (**MUST READ**)
2. [Title](URL) - [Why it matters] (Recommended)

## Actions
- [ ] [Concrete action item]

Step 4: Retention

• Keep last 30 digests
• Notable findings should be captured in MEMORY.md or CLAUDE.md

See Also

• /project-status -- Current project state
• cognitive-core.conf -- Stack configuration