AGENT SKILL

project-bootstrap

Create, modify, inspect, and troubleshoot Solace Agent Mesh project scaffolds without running live services.

License 许可
Apache-2.0
最近更新
2026-09-03
Stars
★ 351
Forks
27

01 前言 Front Matter

name
project-bootstrap
license
Apache 2.0
metadata
{ "disco-role": "operating" }
description
Create, modify, inspect, and troubleshoot Solace Agent Mesh project scaffolds without running live services.
disable-model-invocation
true

02 资源文件 4

03 SKILL.md 正文

Project bootstrap

Use this sub-skill when the task is to create or change a Solace Agent Mesh (SAM) project scaffold with sam init, sam add agent, sam add gateway, or sam add proxy; reason about generated project layout; configure shared services; or dry-inspect generated files.

Do not use this sub-skill for:

  • Plugin packages, plugin installs, plugin catalogs, or plugin builds; route to plugin-lifecycle.
  • Starting projects, brokers, gateways, REST calls, tasks, or Web UI interaction; route to runtime-operations.
  • Authoring workflow DAG internals; route to workflow-authoring.
  • Authoring or running sam eval; route to evaluation.

Fast path

  1. Decide the project creation surface:

    • Browser setup: sam init --gui launches the configuration portal on port 5002.
    • Non-interactive setup: sam init --skip ... uses provided flags or defaults and is suitable for automation.
    • Terminal setup: sam init prompts for browser versus terminal configuration.
  2. For project-level initialization behavior, generated layout, shared config, .env, model provider, broker, Web UI, platform service, artifact, session, and database decisions, use references/project-initialization.md.

  3. For adding components to an existing project, use references/component-scaffolding.md.

  4. Before any live sam run or task submission, run the bundled dry inspector from the generated skill tree:

    python sub-skills/project-bootstrap/scripts/inspect_project.py path/to/sam-project
    

    Add --json for machine-readable output, --strict to fail on warnings, or --self-test to exercise the helper on an embedded tiny fixture.

  5. If the scaffold is incomplete or surprising, map symptoms to fixes in references/troubleshooting.md.

Safety boundaries

  • This sub-skill may parse files, write project scaffolding guidance, and run safe helper-script help/self-test checks.
  • Do not run native repository tests, sam run, broker containers, LLM probes, gateway health checks, task submission, plugin install/build, or evaluation commands from this sub-skill.
  • Treat GUI operations as local configuration-portal bootstrap behavior only; do not assume the browser completed successfully unless generated files or CLI return data prove it.

Minimal scaffold checklist

A typical sam init project contains:

  • requirements.txt pinning the installed SAM package family.
  • .env with namespace, broker, Web UI, platform, logging, authorization, S3/OAuth placeholders, and selected model secrets.
  • configs/shared_config.yaml with broker, optional model anchors, service anchors, data tools, and auto-summarization.
  • configs/logging_config.yaml.
  • configs/agents/main_orchestrator.yaml.
  • configs/gateways/webui.yaml when the Web UI gateway is enabled.
  • configs/services/platform.yaml when the Web UI gateway is enabled.
  • src/__init__.py, plus component source packages created by sam add gateway.

Provenance distilled

This sub-skill was distilled from SAM project, agent, gateway, CLI, installation, configuration, and run-project documentation; CLI init/add implementation; project templates; configuration portal backend behavior; and CLI unit tests covering initialization, add-agent/add-gateway/add-proxy, artifact storage, persistence, GUI handoff, name formatting, overwrite behavior, and error handling. Runtime files are self-contained and do not require opening those source files.

04 同仓库更多 Skills

VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-triton-opt

Make a single PassNet Triton kernel fast AND numerically faithful: performance model (when a replacement can win at all), block/grid/warp tuning, autotune policy, launch overhead, and per-op numeric recipes to pass the dtype baseline tolerances. Use when a pass matches and is correct but speedup below expected, or when correctness fails by small numeric margins.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-solve

END-TO-END playbook for solving one PassNet sample: analyze the computation graphs, decide the optimization strategy (what to fuse, what to replace, what to leave alone), drive the iteration loop, and maximize the sample score. This is the ENTRY skill — invoke it first for any PassNet optimization task; it tells you when to use passnet-pattern-fusion, passnet-triton-opt and passnet-feedback.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-skill

PassNet GPU kernel optimization via compiler passes. Design and implement Triton-based optimization passes, create pass files under ./pass_dir/, self-evaluate with pass_evaluator, and iterate to maximize GPU speedup.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-pattern-fusion

Author PassNet pass files: write patterns that actually MATCH the FX graph, pick fusion regions (which ops to absorb into one Triton kernel), and structure multi-pass submissions with the mandatory shared-dispatch architecture. Use when creating or fixing pattern()/replacement_args()/replacement_func() files, when a pass "failed to match", or when deciding how to fuse multiple kernels/ops.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-orchestrate

Round/batch planning and decision support for PassNet graph optimization. Use for multi-sample triage, worker allocation, eval-budget policy, and post-eval keep/revert decisions. For solving one concrete sample end to end, use passnet-solve as the entry skill; consult this skill only when you need broader planning or a specific decision gate.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-feedback

Fast, deterministic feedback for PassNet work: pre-flight pattern/match verification WITHOUT burning a GPU evaluation, per-node bottleneck analysis of a sample's graphs, and eval-log parsing into per-variant status + estimated score + failure classification. Use BEFORE every GPU evaluation (check_pattern), at the START of a sample (analyze_graph), and AFTER every evaluation (parse_eval_log).

★ 351