All skills
simota avatar

/nexus

@c805268
by shingo imotasimota/agent-skills85 stars
15

Orchestrating multi-specialist task chains and scope-adaptive product delivery: classifies intent, selects and executes the minimum viable chain, aggregates results, and verifies acceptance criteria. For multi-domain tasks, build-first delivery, and product lifecycle execution.

Use this Skill: https://skilld.dev/gh/simota/agent-skills/nexus

This session only. Nothing lands on disk.

referencedelivery-momentum-system.md

≈1.2k tokens on demand. Your agent reads this file only when SKILL.md points to it.

Momentum System

Purpose: Read this file when Nexus[deliver] must prove forward progress, score cycle output, validate velocity, or decide whether Rally is justified.

Contents

  • Forward progress guarantee
  • Weighted progress score
  • Velocity metrics
  • Parallel execution rules
  • Epic prioritization
  • Never idle principle
  • Phase transition momentum

Velocity tracking and forward progress guarantee for Nexus[deliver].


Forward Progress Guarantee

Every execution cycle MUST produce at least one tangible artifact:

Artifact Type Examples
File change Code, config, test, documentation
Test addition New test case, coverage improvement
Decision record Entry in Decision Log
Document generation Spec, diagram, report, README

Zero-Progress Detection: 2 consecutive zero-artifact cycles → Anti-Stall Engine activates. Cycle Definition: One Nexus AUTORUN_FULL chain execution or one Rally team completion.

Weighted Progress Score

Not all artifacts contribute equally to forward progress. Use weighted scoring to distinguish meaningful progress from trivial updates.

Artifact Type Weight Examples
Core Implementation 1.0 Feature code, critical bug fix, API endpoint
Test Addition 0.7 New test suite, significant coverage improvement
Architecture Decision 0.5 ADR entry, schema design, API spec
Configuration Change 0.3 Build config, CI pipeline, environment setup
Documentation Update 0.3 README, API docs, user guide
Decision Log Entry 0.1 Decision record, risk assessment
Minor Update 0.1 Comment addition, formatting, trivial fix

Meaningful Progress Threshold: A cycle's weighted score must reach ≥0.3 to count as progress. Below 0.3, the cycle is treated as zero-progress for Anti-Stall detection purposes.

Example: A cycle that only adds a Decision Log entry (0.1) = zero-progress. A cycle that adds a test (0.7) = meaningful progress. A cycle with config change (0.3) = threshold met.


Velocity Metrics

Metric Calculation Healthy Warning Critical
Epic completion rate completed / planned × 100% ≥70% 50-69% <50%
Cycle time per Epic actual / estimated ≤1.2x 1.2-2x >2x
Stall frequency stalls / total_cycles × 100% ≤10% 10-25% >25%
Agent success rate successful / total_calls × 100% ≥80% 60-79% <60%

Actions: Healthy → continue · Warning → log, increase monitoring · Critical (1 metric) → Sherpa re-decomposition · Critical (2+) → Anti-Stall Engine L2+.


Parallel Execution Rules

Rally Auto-Launch Conditions

Rally is automatically invoked when:

  1. Independent Epics: 2+ Epics with no shared file dependencies
  2. Domain separation: 2+ technical domains (frontend, backend, infra)
  3. Sherpa parallel_group: Sherpa marks tasks as parallelizable
  4. Phase pattern: Phase naturally supports parallelism (HARDEN, VALIDATE, LAUNCH)

Rally Sizing

Phase Typical Teams Reason
BUILD 2-4 Feature parallelism
HARDEN 3 Security + Performance + Quality
VALIDATE 2 E2E tests + Unit tests
LAUNCH 3 Docs + Demos + CI/CD

File Ownership Protocol

When Rally is active, assign explicit file ownership per team. No two teams write to the same file. Shared dependencies resolved before launch. Integration managed by Nexus[deliver] after Rally completion.


Epic Prioritization

Priority Criteria Action
1. Impact/Effort High impact + Low effort Schedule first
2. Unblocking Completes → unblocks 3+ Epics Boost priority
3. Risk Reduction Addresses highest-risk item or validates critical assumption Boost priority
4. User Visibility Shows tangible user-facing progress Boost priority

Quick Wins: Schedule ≥1 quick win at phase start (completable in single Nexus chain, 1-3 agents). Builds momentum and provides early validation.


Never Idle Principle

If primary path is blocked: Check other Epics in phase → Pre-work for next phase → Address tech debt (Sweep/Zen) → Update docs/diagrams → Add test coverage. Never report "waiting" without actively working.


Phase Transition Momentum

At phase boundaries: Begin next phase prep while current wraps up · Run exit criteria checklist systematically · Carry barely-missed items to next phase as high-priority · Log 2-3 sentence retrospective to journal.

Source: SKILL.md on GitHub

3 warnings13d5 checks · Risk MEDIUM
  • Gen Agent Trust Hub13d

    The 'nexus' skill is a comprehensive multi-agent orchestration framework that manages complex task chains. While it incorporates extensive internal guardrails and verification protocols, it explicitly mandates the use of high-risk flags that bypass security permissions to achieve autonomy. It also provides instructions for establishing persistent tasks via cron and GitHub Actions, and utilizes external research tools to fetch content from the web.

  • Socket13d

    1 alert: gptSecurity

  • Snyk13d

    Risk: LOW · No issues

  • Runlayer6mo

    6/22 files flagged

  • ZeroLeaks5mo

    Score: 93/100 · 2 sections analyzed

Signed by skilld at c805268. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 2 days ago.

Activeupdated 2 weeks ago

README badge

README badge for simota/agent-skills/nexus