x0tta6bl4-80-20-executor - SKILL.md Agent Skill

name: x0tta6bl4-80-20-executor description: Run an 80/20 automation cycle for x0tta6bl4 with preconfigured agent profiles and prioritized checks. Use when the user asks to maximize speed and effectiveness, run agents and skills automatically, continue development efficiently, or get fast signal on regressions with minimal compute.

x0tta6bl4 80/20 Executor

Overview

Execute a high-leverage cycle that focuses on the small set of checks giving most value:

MaaS API regression (nodes + escrow + analytics + acl + billing + constants + mesh + metering + swarm_sync + resilience units)
security crypto scan
coverage-gap prioritization

Use this skill for automation-first execution when breadth should be traded for fast actionable signal.

All agent profiles in this skill must apply the shared thinking contract in docs/ai_agents/THINKING_TECHNIQUES_FOR_AGENTS.md before selecting commands, triaging failures, or reporting completion.

Workflow

Choose mode:

auto: pick quick/focused/full from current changed files (default; falls back to focused if git scan times out)
quick: security + coverage only (fastest)
focused: MaaS regression + security + coverage (default manual 80/20)
full: focused + network regression

Built-in per-agent timeouts:

quick: 900s
focused: 2100s
full: 2700s

Contention-aware behavior:

If external pytest processes are detected, the runner auto-increases timeout (up to +900s) and reduces parallelism by 1.

Run the cycle script:

skills/x0tta6bl4-80-20-executor/scripts/run_80_20_cycle.sh auto

Read generated run summary from scripts/agents/run_agent_cycle.py output:

.../summary.md for operator view
.../summary.json for machine-readable post-processing

Prioritize fixes from blocking failures first, then rerun in focused mode.

Commands

Quick signal:

skills/x0tta6bl4-80-20-executor/scripts/run_80_20_cycle.sh quick

Auto-selected cycle:

skills/x0tta6bl4-80-20-executor/scripts/run_80_20_cycle.sh auto

Default manual 80/20 cycle:

skills/x0tta6bl4-80-20-executor/scripts/run_80_20_cycle.sh focused

Full cycle:

skills/x0tta6bl4-80-20-executor/scripts/run_80_20_cycle.sh full

Pass extra runner flags through to the agent cycle:

skills/x0tta6bl4-80-20-executor/scripts/run_80_20_cycle.sh focused --strict

Optional modular billing sanity check:

PYTEST_DISABLE_PLUGIN_AUTOLOAD=1 python3 -m pytest -q -o addopts='' --confcutdir=/mnt/projects/tests/unit/api tests/unit/api/test_maas_services_billing_unit.py

References

references/agent_cycle_profiles_80_20.json: curated commands and mapped skills per agent.