hydra-deep-research-self-healing
Run deep-research loops with HYDRA/Agent Relay using self-healing workflow methods (sense, plan, execute, verify, recover), combining arXiv routing, Playwright evidence capture, synthesis smoke checks, and CI triage hooks. Use when operating long-running 24/7 research automation with deterministic cross-talk logging.
Best use case
hydra-deep-research-self-healing is best used when you need a repeatable AI agent workflow instead of a one-off prompt.
Run deep-research loops with HYDRA/Agent Relay using self-healing workflow methods (sense, plan, execute, verify, recover), combining arXiv routing, Playwright evidence capture, synthesis smoke checks, and CI triage hooks. Use when operating long-running 24/7 research automation with deterministic cross-talk logging.
Teams using hydra-deep-research-self-healing should expect a more consistent output, faster repeated execution, less prompt rewriting.
When to use this skill
- You want a reusable workflow that can be run more than once with consistent structure.
When not to use this skill
- You only need a quick one-off answer and do not need a reusable workflow.
- You cannot install or maintain the underlying files, dependencies, or repository context.
Installation
Claude Code / Cursor / Codex
Manual Installation
- Download SKILL.md from GitHub
- Place it in
.claude/skills/hydra-deep-research-self-healing/SKILL.mdinside your project - Restart your AI agent — it will auto-discover the skill
How hydra-deep-research-self-healing Compares
| Feature / Agent | hydra-deep-research-self-healing | Standard Approach |
|---|---|---|
| Platform Support | Not specified | Limited / Varies |
| Context Awareness | High | Baseline |
| Installation Complexity | Unknown | N/A |
Frequently Asked Questions
What does this skill do?
Run deep-research loops with HYDRA/Agent Relay using self-healing workflow methods (sense, plan, execute, verify, recover), combining arXiv routing, Playwright evidence capture, synthesis smoke checks, and CI triage hooks. Use when operating long-running 24/7 research automation with deterministic cross-talk logging.
Where can I find the source code?
You can find the source code on GitHub using the link provided at the top of the page.
Related Guides
SKILL.md Source
# Hydra Deep Research Self-Healing Use this skill for terminal-first research operations that must run continuously and recover from failures. ## Loop Stages 1. `SENSE` - Route arXiv research lane. ```powershell python C:\Users\issda\SCBE-AETHERMOORE\scripts\system\browser_chain_dispatcher.py --domain arxiv.org --task research --engine playwriter ``` - Capture deterministic page evidence. ```powershell python C:\Users\issda\SCBE-AETHERMOORE\scripts\system\playwriter_lane_runner.py --session 1 --task title python C:\Users\issda\SCBE-AETHERMOORE\scripts\system\playwriter_lane_runner.py --session 1 --task snapshot ``` 2. `PLAN` - Run synthesis planner packet. ```powershell powershell -NoProfile -ExecutionPolicy Bypass -File .\scripts\system\run_hydra_agent-relay_synthesis.ps1 ``` 3. `EXECUTE` - Run stack readiness + External Agent lane. ```powershell powershell -NoProfile -ExecutionPolicy Bypass -File .\scripts\system\run_hydra_research.ps1 -All ``` 4. `VERIFY` - Run full-system smoke and export summary. ```powershell powershell -NoProfile -ExecutionPolicy Bypass -File .\scripts\system\run_synthesis_pipeline.ps1 ``` 5. `RECOVER` - Start watchdog for self-healing restart loop. ```powershell powershell -NoProfile -ExecutionPolicy Bypass -File .\scripts\system\bootstrap_24x7_local.ps1 -RunWatchdogNow -StartWatchdogLoop -WatchdogIntervalMinutes 5 ``` ## CI Triage Hook When research code is blocked by failing GitHub checks, route to CI triage lane: ```powershell python C:\Users\issda\.codex\skills\gh-fix-ci\scripts\inspect_pr_checks.py --repo "C:\Users\issda\SCBE-AETHERMOORE" ``` ## Cross-Talk Contract Emit packet each state transition (`start`, `blocked`, `done`) with where/why/how metadata: ```powershell & C:\Users\issda\SCBE-AETHERMOORE\scripts\system\terminal_crosstalk_emit.ps1 -TaskId "DEEP-RESEARCH-LOOP" -Summary "Loop update" -Status "in_progress" -Where "hydra-loop" -Why "continuous governed research" -How "sense-plan-execute-verify-recover" ``` ## 24/7 VM Operations Use VM control wrapper for remote operation: ```powershell & C:\Users\issda\SCBE-AETHERMOORE\scripts\system\vm_agent-relay_24x7.ps1 -Action status -VmHost <VM_IP> -User ubuntu -KeyPath <SSH_KEY> ``` Run deep-research loop continuously: ```powershell powershell -NoProfile -ExecutionPolicy Bypass -File C:\Users\issda\SCBE-AETHERMOORE\scripts\system\run_deep_research_self_healing.ps1 -Topic "your topic" -Continuous -SleepSeconds 120 -UsePlaywriter -RunCiTriage ``` ## Safety Gate: Onion/Tor - Do not run automated Tor or dark-web crawling/mapping from this skill. - If onion sources are needed for compliance or threat-intel validation, require explicit legal approval and perform manual, narrowly scoped checks outside autonomous loops. - Keep autonomous verification on public-web + first-party sources + archived snapshots.
Related Skills
scbe-research-training-bridge
Stage arXiv evidence and Obsidian markdown into source-grounded Hugging Face training bundles for research, review, and later SFT runs.
scbe-self-improvement-skill-management
Run the SCBE self-improvement loop by documenting completed work, creating new skills, updating existing skills, and propagating SCBE knowledge to other AI systems via context documents and structured knowledge files. Use when users ask to create/update skills, formalize repeated SCBE workflows, enforce skill approval gates, generate transfer prompts for Claude/Gemini/custom agents, or publish SCBE knowledge artifacts to Hugging Face datasets.
scbe-research-publishing-autopilot
Build and operate multi-hour SCBE research-to-publishing loops for articles and social posts with retrigger logic, evidence gates, and performance monitoring. Use when repeatable growth execution must stay lore/code accurate and avoid invention misrepresentation.
scbe-manhwa-video-picture-research
Research manhwa recap video and image workflows and emit story-first, source-traceable scene packets for production, narration, and publishing.
scbe-experimental-research-safe-integration
Design and execute advanced experimental research with fail-closed safety integration for SCBE systems. Use when users ask to test novel kernels/manifolds, map proposals into the 21D/M4 state layout, add semantic mesh overlays (including 230-bit envelopes), or move experimental results toward production safely.
scbe-ai2ai-kindle-research
Run Kindle-focused AI2AI research/debate workflows through n8n and SCBE bridge endpoints with governance thresholds.
notion-research-documentation
Research across Notion and synthesize into structured documentation; use when gathering info from multiple Notion sources to produce briefs, comparisons, or reports with citations.
hydra-node-terminal-browsing
Use when HYDRA tasks need Node.js-first terminal browsing, deterministic page extraction, and evidence JSON output that can be routed into hydra swarm workflows.
scbe-web-research-verified
Run verifiable web research across mainstream news, niche outlets, and academic primary sources, then return dated evidence with confidence scoring. Use when asked to research current events, science claims, market context, or source-grounded recommendations.
scbe-training-pair-authoring
Create prompt and response and metadata training pairs from SCBE documents, repair traces, terminal sessions, and operational workflows using the repository's canonical dataset contract and provenance rules.
scbe-spin-conversation-engine
Generate SFT training data via radial matrix conversation pivots with D&D-style combat research mode. Produces diverse, cost-effective training pairs with Sacred Tongue encoding, golden spiral problem distribution, and harmonic re-attunement.
scbe-document-management
Consolidate overlapping docs, classify files by authority, and keep SCBE repo documents aligned with runtime truth. Use when the repo has drift between canonical docs, public docs, proposal notes, research branches, and generated evidence.