// Extract system improvements and monitor Anthropic ecosystem. USE WHEN upgrade, check Anthropic, new Claude features.
| name | PAIUpgrade |
| description | Extract system improvements and monitor Anthropic ecosystem. USE WHEN upgrade, check Anthropic, new Claude features. |
| context | fork |
Before executing, check for user customizations at:
~/.opencode/PAI/USER/SKILLCUSTOMIZATIONS/PAIUpgrade/
If this directory exists, load and apply any PREFERENCES.md, configurations, or resources found there. These override default behavior. If the directory does not exist, proceed with skill defaults.
Universal system upgrade skill with two modes:
When executing a workflow, do BOTH:
Send voice notification:
curl -s -X POST http://localhost:8888/notify \
-H "Content-Type: application/json" \
-d '{"message": "Running the WORKFLOWNAME workflow from the PAIUpgrade skill"}' \
> /dev/null 2>&1 &
Output text notification:
Running the **WorkflowName** workflow from the **PAIUpgrade** skill...
Route to the appropriate workflow based on the request.
When executing a workflow, output this notification directly:
Running the **WorkflowName** workflow from the **PAIUpgrade** skill...
| Workflow | Trigger | File |
|---|---|---|
| CheckForUpgrades | "check for upgrades", "check sources", "any updates", "check Anthropic", "check YouTube" | Workflows/CheckForUpgrades.md |
| ResearchUpgrade | "research this upgrade", "deep dive on [feature]", "further research" | Workflows/ResearchUpgrade.md |
| ReleaseNotesDeepDive | "analyze release notes", "deep dive release" | Workflows/ReleaseNotesDeepDive.md |
| FindSources | "find upgrade sources", "find new sources", "discover channels" | Workflows/FindSources.md |
Universal Input -> System Upgrade Recommendations
Takes ANY content type and performs deep thinking analysis to extract insights and identify concrete system infrastructure improvement opportunities.
Analyzes content across 10 dimensions:
"No Gaps Found" is a VALID and often CORRECT output.
If analysis shows the system already implements everything in the content:
Only if genuine gaps exist, output prioritized recommendations:
What is NOT a valid recommendation:
These are busywork, not upgrades. If the system does it, we don't need to "document" it as an upgrade.
Proactive ecosystem monitoring for PAI-relevant updates
Sources Monitored:
Tool: Tools/Anthropic.ts
YouTube channels are configured via the Skill Customization Layer.
See ~/.opencode/PAI/USER/SKILLCUSTOMIZATIONS/PAIUpgrade/ for user-specific channels.
Features:
| Tool | Purpose |
|---|---|
Tools/Anthropic.ts | Check Anthropic sources for updates |
Skill Files:
sources.json - Anthropic sources config (30+ sources)youtube-channels.json - Base YouTube channels (empty - uses customization)State/last-check.json - Anthropic stateState/youtube-videos.json - YouTube stateUser Customizations (~/.opencode/PAI/USER/SKILLCUSTOMIZATIONS/PAIUpgrade/):
EXTEND.yaml - Extension manifestyoutube-channels.json - User's personal YouTube channelsUse bun ~/.opencode/PAI/Tools/LoadSkillConfig.ts to load configs with customizations merged.
The skill has four complementary workflows:
| Workflow | Purpose |
|---|---|
| CheckForUpgrades | Monitor configured sources (Anthropic + YouTube) for new content |
| ResearchUpgrade | Deep dive on discovered features to understand implementation |
| ReleaseNotesDeepDive | Specialized research on Claude Code release notes |
| FindSources | Discover and evaluate new sources to add to monitoring |
Typical flow:
Identifies combinations of improvements that multiply value:
When analyzing multiple pieces of content over time:
Compares content insights to the system's current capabilities:
The skill improves its own recommendations over time:
parser:
research:
be-creative:
development:
VideoTranscript:
History Capture:
~/.opencode/History/research/YYYY-MM/Todo System:
Agent Delegation:
Example 1: Check for upgrades
User: "check for upgrades"
→ Invokes CheckForUpgrades workflow
→ Runs Anthropic.ts tool (30+ sources)
→ Checks YouTube channels (from USER config)
→ Combines into prioritized upgrade report
Example 2: Research a discovered feature
User: "research the new context forking feature"
→ Invokes ResearchUpgrade workflow
→ Spawns parallel research agents
→ Searches GitHub, docs, blog for details
→ Maps to PAI architecture opportunities
→ Outputs implementation recommendations
Example 3: Deep dive on release notes
User: "deep dive the latest release notes"
→ Invokes ReleaseNotesDeepDive workflow
→ Runs /release-notes to capture features
→ Launches parallel research agents for each feature
→ Maps to PAI architecture opportunities
→ Outputs prioritized upgrade roadmap with citations
Example 4: Find new sources
User: "find new upgrade sources"
→ Invokes FindSources workflow
→ Searches for relevant YouTube channels
→ Evaluates and ranks findings
→ Outputs recommendations with add instructions
Every recommendation must have:
Avoid:
This skill embodies the system's commitment to continuous improvement and learning from the broader ecosystem while maintaining our architectural principles and preferences.
Personal AI Infrastructure core. The authoritative reference for how PAI works.
Life OS and project management. USE WHEN life goals, projects, dependencies, TELOS, books, movies, tracking.
Security assessment and intelligence. USE WHEN recon, reconnaissance, port scan, subdomain, DNS, WHOIS, web assessment, pentest, vulnerability, security scan, prompt injection, jailbreak, LLM security, security news, breaches, annual reports, threat landscape.
Web security assessment. USE WHEN web assessment, pentest, security testing, vulnerability scan. SkillSearch('webassessment') for docs.
PowerPoint processing. USE WHEN pptx, PowerPoint, slides. SkillSearch('pptx') for docs.
Utility and helper skills. USE WHEN aphorisms, quotes, browser automation, Cloudflare, create CLI, build CLI, create skill, process documents, PDF, Word, Excel, evaluations, evals, fabric patterns, PAI upgrade, parser, prompting, templates.