Writing Skills
writing-skills
效率与协作技能,来自 obra/superpowers,查看原文获取完整用途和操作说明。
config-supervisor-deployment【部署与运维】Use this quip-miner sub-skill for TOML configuration, production supervisor mode, mode resolution, Docker/container deployment, telemetry bind settings, and safe deployment troubleshooting.
独立 GitHub 快照 · 自动分类 · 以原作者文档为准
Use this sub-skill when the user needs to create, migrate, lint, or explain a quip-miner TOML config; choose direct subcommands vs production supervisor mode; run a Docker/container deployment; or diagnose config-mode conflicts.
references/config-schema.md for [miner], [submission], backend inventory tables, validators, signer key, telemetry, aliases, and mempool placement.references/supervisor-and-deployment.md for quip-miner --config, --mode narrowing, Docker /data/config.toml, shared memory, entrypoint behavior, and systemd cautions.references/troubleshooting.md for malformed TOML, invalid [miner] mempool, backend conflict errors, invalid validators, and submission tip/retry validation.scripts/render_config_template.py to draft a safe CPU/CUDA/QPU TOML skeleton, then use the root scripts/quip_config_lint.py against the active installed package.quip-miner --config config.toml; the supervisor starts one child per declared backend group and a telemetry aggregator when REST is enabled.quip-miner cpu, gpu, qpu) accept CLI flags for focused runs, but if the TOML already declares a backend group, conflicting CLI inventory flags fail fast.--mode cpu|gpu|qpu narrows a multi-backend config at runtime. It is CLI-only and must not be written into TOML.[miner] is for connection, signer, faucet, telemetry, identity, logging, and global submission/reward thresholds. Hardware inventory lives in top-level backend sections.mempool belongs inside backend sections, not [miner]. CPU/GPU default on; QPU default off.../identity-wallet-bootstrap/SKILL.md.../mining-backends/SKILL.md.../telemetry-attempt-archive/SKILL.md.../maintainer-testing-release/SKILL.md.python scripts/render_config_template.py --backend cpu --num-cpus 4
python scripts/render_config_template.py --backend cuda --cuda-devices 0,1 --rest-port 8086
python scripts/quip_config_lint.py --config config.toml --json
quip-miner resolve-modes --config config.toml
Do not start live miners solely to validate TOML shape; parser and mode-resolution checks are safer first.