From e5e43343cca8aef2c4f136a912dc7c75ab5b37e5 Mon Sep 17 00:00:00 2001 From: Wang Siyuan Date: Tue, 3 Mar 2026 22:32:26 +0800 Subject: [PATCH 1/7] docs(skills): add keepgpu repository workflow skill --- README.md | 1 + .../skillcheck-free-report.md | 23 +++++ skills/keepgpu-repo-workflow/SKILL.md | 94 +++++++++++++++++++ .../keepgpu-repo-workflow/agents/openai.yaml | 3 + 4 files changed, 121 insertions(+) create mode 100644 docs/skills/keepgpu-repo-workflow/skillcheck-free-report.md create mode 100644 skills/keepgpu-repo-workflow/SKILL.md create mode 100644 skills/keepgpu-repo-workflow/agents/openai.yaml diff --git a/README.md b/README.md index 3dcceb7..2dddc3e 100644 --- a/README.md +++ b/README.md @@ -5,6 +5,7 @@ [![DOI](https://zenodo.org/badge/987167271.svg)](https://doi.org/10.5281/zenodo.17129114) [![Ask DeepWiki](https://deepwiki.com/badge.svg)](https://deepwiki.com/Wangmerlyn/KeepGPU) [![CodeRabbit Pull Request Reviews](https://img.shields.io/coderabbit/prs/github/Wangmerlyn/KeepGPU)](https://coderabbit.ai/dashboard/gh/Wangmerlyn/KeepGPU) +[![SkillCheck Passed](https://raw.githubusercontent.com/olgasafonova/skillcheck-free/main/skill-check/passed.svg)](https://github.com/olgasafonova/skillcheck-free) **Keep GPU** keeps shared GPUs from being reclaimed while you prep data, debug, or coordinate multi-stage pipelines. It allocates just enough VRAM and issues lightweight CUDA work so schedulers observe an “active” device—without running a full training job. diff --git a/docs/skills/keepgpu-repo-workflow/skillcheck-free-report.md b/docs/skills/keepgpu-repo-workflow/skillcheck-free-report.md new file mode 100644 index 0000000..d68b226 --- /dev/null +++ b/docs/skills/keepgpu-repo-workflow/skillcheck-free-report.md @@ -0,0 +1,23 @@ +# SkillCheck Free Report: keepgpu-repo-workflow + +Validation method: applied the SkillCheck Free rule set from `skill-check/SKILL.md` in . + +## Summary + +- Critical: 0 +- Warnings: 0 +- Suggestions: 0 + +## Strengths detected + +- Description includes activation triggers (8.3) +- Description includes negative triggers (8.7) +- Skill specifies output format with examples (8.4) +- Skill includes example section (8.1) +- Skill documents error handling or limitations (8.2) +- Skill uses structured instructions (8.5) +- Skill documents prerequisites (8.6) + +## Checked files + +- `skills/keepgpu-repo-workflow/SKILL.md` diff --git a/skills/keepgpu-repo-workflow/SKILL.md b/skills/keepgpu-repo-workflow/SKILL.md new file mode 100644 index 0000000..899e45e --- /dev/null +++ b/skills/keepgpu-repo-workflow/SKILL.md @@ -0,0 +1,94 @@ +--- +name: keepgpu-repo-workflow +description: Implement, test, and document changes in the KeepGPU repository while preserving CLI, Python API, and MCP parity, platform boundaries, and GPU-safe behavior on no-GPU CI runners. Use when tasks request code changes, bug fixes, refactors, tests, docs updates, releases, or pull-request work in this repository; do not use for unrelated repositories or generic Python questions. +--- + +# KeepGPU Repository Workflow + +Follow this workflow to make reliable, review-ready changes in KeepGPU. + +## Prerequisites + +- Work from repository root. +- Confirm branch starts from latest `main`. +- Keep user-facing text and comments in English. + +## Repository map + +- CLI entrypoint: `src/keep_gpu/cli.py` +- Single-GPU controllers: `src/keep_gpu/single_gpu_controller/` +- Global controller orchestration: `src/keep_gpu/global_gpu_controller/` +- MCP server: `src/keep_gpu/mcp/server.py` +- Platform probing: `src/keep_gpu/utilities/platform_manager.py` +- GPU telemetry helpers: `src/keep_gpu/utilities/gpu_info.py` + +## Implementation rules + +1. Start from a new branch and keep diffs focused. +2. For non-trivial work, create or update a plan in `docs/plans/` with background, goal, solution, and todo items. +3. Keep platform detection centralized in `platform_manager.py`; do not spread platform branching across unrelated modules. +4. Keep telemetry logic in `gpu_info.py` or related utility modules. +5. Preserve controller flow: global controller orchestrates per-GPU controllers; single-GPU controllers handle keep/release loops. +6. Keep CUDA telemetry on `nvidia-ml-py` (`pynvml` module import) and keep ROCm support optional with graceful failure on non-ROCm hosts. +7. Update docs when behavior changes for CLI flags, controllers, platform support, or MCP methods. +8. Keep commits narrow and use `type(scope): summary` commit messages. + +## Validation order + +Run the smallest relevant checks first, then broader checks. + +1. Targeted tests for changed modules. +2. Broader tests if changes touch shared logic. +3. `pre-commit run --all-files` before push. +4. `mkdocs build` when docs are changed. + +Preferred targeted test commands: + +```bash +pytest tests/cuda_controller tests/global_controller tests/utilities/test_platform_manager.py +pytest tests -k threshold +pytest tests/mcp tests/utilities/test_gpu_info.py +``` + +ROCm-only command (run only on ROCm-capable machines): + +```bash +pytest --run-rocm tests/rocm_controller +``` + +## Output format + +Return results in this structure: + +```markdown +## KeepGPU Task Result + +### Changes +- + +### Validation +- Targeted tests: +- Broader checks: +- Docs build: +- Pre-commit: + +### Risks +- +``` + +## Example + +User request: "Add a new CLI flag for busy-threshold defaults and update docs." + +Execution pattern: + +1. Update `src/keep_gpu/cli.py` and related controller wiring. +2. Add or update tests in `tests/test_cli_thresholds.py` and relevant controller tests. +3. Update user docs and README sections for the flag behavior. +4. Run targeted tests, then `pre-commit run --all-files`, then `mkdocs build`. +5. Prepare a focused commit and PR summary. + +## Limitations + +- GPU hardware is often unavailable in CI; guard hardware-dependent logic and tests. +- Do not bump versions, retag releases, or alter release metadata unless explicitly requested. diff --git a/skills/keepgpu-repo-workflow/agents/openai.yaml b/skills/keepgpu-repo-workflow/agents/openai.yaml new file mode 100644 index 0000000..805a9e1 --- /dev/null +++ b/skills/keepgpu-repo-workflow/agents/openai.yaml @@ -0,0 +1,3 @@ +display_name: KeepGPU Repo Workflow +short_description: Implement and validate KeepGPU repository changes with architecture and testing guardrails. +default_prompt: Use this skill to implement, test, and document changes in KeepGPU while preserving CLI/API/MCP parity and platform-specific boundaries. From b95170d9593dc4604c067330d985ae42923135aa Mon Sep 17 00:00:00 2001 From: Wang Siyuan Date: Tue, 3 Mar 2026 22:51:05 +0800 Subject: [PATCH 2/7] docs(skills): refocus keepgpu skill on CLI usage --- .../skillcheck-free-report.md | 2 +- skills/keepgpu-repo-workflow/SKILL.md | 149 +++++++++++------- .../keepgpu-repo-workflow/agents/openai.yaml | 6 +- 3 files changed, 99 insertions(+), 58 deletions(-) diff --git a/docs/skills/keepgpu-repo-workflow/skillcheck-free-report.md b/docs/skills/keepgpu-repo-workflow/skillcheck-free-report.md index d68b226..39d67d2 100644 --- a/docs/skills/keepgpu-repo-workflow/skillcheck-free-report.md +++ b/docs/skills/keepgpu-repo-workflow/skillcheck-free-report.md @@ -1,6 +1,6 @@ # SkillCheck Free Report: keepgpu-repo-workflow -Validation method: applied the SkillCheck Free rule set from `skill-check/SKILL.md` in . +Validation method: applied the SkillCheck Free rule set from `skill-check/SKILL.md` in after updating the skill to focus on KeepGPU CLI usage and installation instructions. ## Summary diff --git a/skills/keepgpu-repo-workflow/SKILL.md b/skills/keepgpu-repo-workflow/SKILL.md index 899e45e..643cde0 100644 --- a/skills/keepgpu-repo-workflow/SKILL.md +++ b/skills/keepgpu-repo-workflow/SKILL.md @@ -1,94 +1,135 @@ --- name: keepgpu-repo-workflow -description: Implement, test, and document changes in the KeepGPU repository while preserving CLI, Python API, and MCP parity, platform boundaries, and GPU-safe behavior on no-GPU CI runners. Use when tasks request code changes, bug fixes, refactors, tests, docs updates, releases, or pull-request work in this repository; do not use for unrelated repositories or generic Python questions. +description: Install and operate the KeepGPU CLI to keep reserved GPUs active during data prep, debugging, and orchestration downtime. Use when users ask for keep-gpu command construction, tuning (--vram, --interval, --busy-threshold), installation from this repository, or runtime troubleshooting of keep-gpu sessions; do not use for repository development, code refactoring, or unrelated Python tooling. --- -# KeepGPU Repository Workflow +# KeepGPU CLI Operator -Follow this workflow to make reliable, review-ready changes in KeepGPU. +Use this workflow to run `keep-gpu` safely and effectively. ## Prerequisites -- Work from repository root. -- Confirm branch starts from latest `main`. -- Keep user-facing text and comments in English. +- Confirm at least one GPU is visible (`python -c "import torch; print(torch.cuda.device_count())"`). +- Run commands in a shell where CUDA/ROCm drivers are already available. +- Use `Ctrl+C` to stop KeepGPU and release memory cleanly. -## Repository map +## Install KeepGPU -- CLI entrypoint: `src/keep_gpu/cli.py` -- Single-GPU controllers: `src/keep_gpu/single_gpu_controller/` -- Global controller orchestration: `src/keep_gpu/global_gpu_controller/` -- MCP server: `src/keep_gpu/mcp/server.py` -- Platform probing: `src/keep_gpu/utilities/platform_manager.py` -- GPU telemetry helpers: `src/keep_gpu/utilities/gpu_info.py` +Install PyTorch first for your platform, then install KeepGPU. -## Implementation rules +### Option A: Install from package index -1. Start from a new branch and keep diffs focused. -2. For non-trivial work, create or update a plan in `docs/plans/` with background, goal, solution, and todo items. -3. Keep platform detection centralized in `platform_manager.py`; do not spread platform branching across unrelated modules. -4. Keep telemetry logic in `gpu_info.py` or related utility modules. -5. Preserve controller flow: global controller orchestrates per-GPU controllers; single-GPU controllers handle keep/release loops. -6. Keep CUDA telemetry on `nvidia-ml-py` (`pynvml` module import) and keep ROCm support optional with graceful failure on non-ROCm hosts. -7. Update docs when behavior changes for CLI flags, controllers, platform support, or MCP methods. -8. Keep commits narrow and use `type(scope): summary` commit messages. +```bash +# CUDA example (change cu121 to your CUDA version) +pip install --index-url https://download.pytorch.org/whl/cu121 torch +pip install keep-gpu +``` -## Validation order +```bash +# ROCm example (change rocm6.1 to your ROCm version) +pip install --index-url https://download.pytorch.org/whl/rocm6.1 torch +pip install keep-gpu[rocm] +``` -Run the smallest relevant checks first, then broader checks. +### Option B: Install from this repository -1. Targeted tests for changed modules. -2. Broader tests if changes touch shared logic. -3. `pre-commit run --all-files` before push. -4. `mkdocs build` when docs are changed. +```bash +pip install -e . +``` -Preferred targeted test commands: +For ROCm users from repo checkout: ```bash -pytest tests/cuda_controller tests/global_controller tests/utilities/test_platform_manager.py -pytest tests -k threshold -pytest tests/mcp tests/utilities/test_gpu_info.py +pip install -e ".[rocm]" ``` -ROCm-only command (run only on ROCm-capable machines): +Verify installation: ```bash -pytest --run-rocm tests/rocm_controller +keep-gpu --help ``` +## Command model + +CLI options to tune: + +- `--gpu-ids`: comma-separated IDs (`0`, `0,1`). If omitted, KeepGPU uses all visible GPUs. +- `--vram`: VRAM to hold per GPU (`512MB`, `1GiB`, or raw bytes). +- `--interval`: seconds between keep-alive cycles. +- `--busy-threshold` (`--util-threshold` alias): if utilization is above this percent, KeepGPU backs off. + +Legacy compatibility: + +- `--threshold` is deprecated but still accepted. +- Numeric `--threshold` maps to busy threshold. +- String `--threshold` maps to VRAM. + +## Agent workflow + +1. Collect workload intent: target GPU IDs, expected hold duration, and whether node is shared. +2. Choose safe defaults when unspecified: `--vram 1GiB`, `--interval 60-120`, `--busy-threshold 25` for shared nodes. +3. Build one concrete command. +4. Provide stop instruction (`Ctrl+C`) and a verification step. +5. If command fails, provide one minimal troubleshooting command at a time. + +## Command templates + +Single GPU while preprocessing: + +```bash +keep-gpu --gpu-ids 0 --vram 1GiB --interval 60 --busy-threshold 25 +``` + +All visible GPUs with lighter load: + +```bash +keep-gpu --vram 512MB --interval 180 +``` + +Background execution on remote shell: + +```bash +nohup keep-gpu --gpu-ids 0 --vram 1GiB --interval 300 > keepgpu.log 2>&1 & +``` + +## Troubleshooting + +- Invalid `--gpu-ids`: ensure comma-separated integers only. +- Allocation failure / OOM: reduce `--vram` or free memory first. +- No utilization telemetry: ensure `nvidia-ml-py` works and `nvidia-smi` is available. +- No GPUs detected: verify drivers, CUDA/ROCm runtime, and `torch.cuda.device_count()`. + ## Output format -Return results in this structure: +Return guidance in this structure: ```markdown -## KeepGPU Task Result +## KeepGPU CLI Plan + +### Install +- -### Changes -- +### Run Command +- -### Validation -- Targeted tests: -- Broader checks: -- Docs build: -- Pre-commit: +### Verify +- -### Risks -- +### Stop +- Press Ctrl+C (or kill background PID) to release VRAM. ``` ## Example -User request: "Add a new CLI flag for busy-threshold defaults and update docs." +User request: "Install from this repo and keep GPU 0 alive while I preprocess." -Execution pattern: +Suggested response shape: -1. Update `src/keep_gpu/cli.py` and related controller wiring. -2. Add or update tests in `tests/test_cli_thresholds.py` and relevant controller tests. -3. Update user docs and README sections for the flag behavior. -4. Run targeted tests, then `pre-commit run --all-files`, then `mkdocs build`. -5. Prepare a focused commit and PR summary. +1. Install: `pip install -e .` +2. Run: `keep-gpu --gpu-ids 0 --vram 1GiB --interval 60 --busy-threshold 25` +3. Verify: check CLI logs for keep loop activity; stop with `Ctrl+C` when done. ## Limitations -- GPU hardware is often unavailable in CI; guard hardware-dependent logic and tests. -- Do not bump versions, retag releases, or alter release metadata unless explicitly requested. +- KeepGPU is not a scheduler; it only keeps already accessible GPUs active. +- KeepGPU behavior depends on cluster policy; some schedulers require higher VRAM or tighter intervals. diff --git a/skills/keepgpu-repo-workflow/agents/openai.yaml b/skills/keepgpu-repo-workflow/agents/openai.yaml index 805a9e1..67fbe49 100644 --- a/skills/keepgpu-repo-workflow/agents/openai.yaml +++ b/skills/keepgpu-repo-workflow/agents/openai.yaml @@ -1,3 +1,3 @@ -display_name: KeepGPU Repo Workflow -short_description: Implement and validate KeepGPU repository changes with architecture and testing guardrails. -default_prompt: Use this skill to implement, test, and document changes in KeepGPU while preserving CLI/API/MCP parity and platform-specific boundaries. +display_name: KeepGPU CLI Operator +short_description: Install and operate keep-gpu CLI commands safely on shared GPU machines. +default_prompt: Use this skill to install KeepGPU and build run-ready keep-gpu commands with sensible defaults, verification steps, and troubleshooting. From dfd85306a0427fb6e7cf66f57018c3dbaa02f270 Mon Sep 17 00:00:00 2001 From: Wang Siyuan Date: Wed, 4 Mar 2026 12:52:00 +0800 Subject: [PATCH 3/7] docs(skills): clarify source checkout install path --- skills/keepgpu-repo-workflow/SKILL.md | 16 +++++++++++++--- 1 file changed, 13 insertions(+), 3 deletions(-) diff --git a/skills/keepgpu-repo-workflow/SKILL.md b/skills/keepgpu-repo-workflow/SKILL.md index 643cde0..5c66c04 100644 --- a/skills/keepgpu-repo-workflow/SKILL.md +++ b/skills/keepgpu-repo-workflow/SKILL.md @@ -31,12 +31,22 @@ pip install --index-url https://download.pytorch.org/whl/rocm6.1 torch pip install keep-gpu[rocm] ``` -### Option B: Install from this repository +### Option B: Install from a KeepGPU source checkout (explicit path) + +Use this option when KeepGPU is checked out locally. Do not run this from the skill folder itself. ```bash +git clone https://github.com/Wangmerlyn/KeepGPU.git +cd KeepGPU pip install -e . ``` +If the checkout already exists somewhere else, install by absolute path: + +```bash +pip install -e /absolute/path/to/KeepGPU +``` + For ROCm users from repo checkout: ```bash @@ -121,11 +131,11 @@ Return guidance in this structure: ## Example -User request: "Install from this repo and keep GPU 0 alive while I preprocess." +User request: "Install from a local KeepGPU checkout and keep GPU 0 alive while I preprocess." Suggested response shape: -1. Install: `pip install -e .` +1. Install: `pip install -e /absolute/path/to/KeepGPU` 2. Run: `keep-gpu --gpu-ids 0 --vram 1GiB --interval 60 --busy-threshold 25` 3. Verify: check CLI logs for keep loop activity; stop with `Ctrl+C` when done. From 22c79ff6d86dbd726ec624acb111b8d2e9ecd0a1 Mon Sep 17 00:00:00 2001 From: Wang Siyuan Date: Wed, 4 Mar 2026 12:54:15 +0800 Subject: [PATCH 4/7] docs(skills): prefer git URL install path --- skills/keepgpu-repo-workflow/SKILL.md | 30 ++++++++++++++++++++++----- 1 file changed, 25 insertions(+), 5 deletions(-) diff --git a/skills/keepgpu-repo-workflow/SKILL.md b/skills/keepgpu-repo-workflow/SKILL.md index 5c66c04..76c4f0b 100644 --- a/skills/keepgpu-repo-workflow/SKILL.md +++ b/skills/keepgpu-repo-workflow/SKILL.md @@ -31,9 +31,29 @@ pip install --index-url https://download.pytorch.org/whl/rocm6.1 torch pip install keep-gpu[rocm] ``` -### Option B: Install from a KeepGPU source checkout (explicit path) +### Option B: Install directly from Git URL (no local clone) -Use this option when KeepGPU is checked out locally. Do not run this from the skill folder itself. +Prefer this option when users only need the CLI and do not need local source edits. This avoids checkout directory and cleanup overhead. + +```bash +pip install "git+https://github.com/Wangmerlyn/KeepGPU.git" +``` + +If SSH access is configured: + +```bash +pip install "git+ssh://git@github.com/Wangmerlyn/KeepGPU.git" +``` + +ROCm variant from Git URL: + +```bash +pip install "keep_gpu[rocm] @ git+https://github.com/Wangmerlyn/KeepGPU.git" +``` + +### Option C: Install from a local source checkout (explicit path) + +Use this option only when users already have a local checkout or plan to edit source. ```bash git clone https://github.com/Wangmerlyn/KeepGPU.git @@ -47,7 +67,7 @@ If the checkout already exists somewhere else, install by absolute path: pip install -e /absolute/path/to/KeepGPU ``` -For ROCm users from repo checkout: +For ROCm users from local checkout: ```bash pip install -e ".[rocm]" @@ -131,11 +151,11 @@ Return guidance in this structure: ## Example -User request: "Install from a local KeepGPU checkout and keep GPU 0 alive while I preprocess." +User request: "Install KeepGPU from GitHub and keep GPU 0 alive while I preprocess." Suggested response shape: -1. Install: `pip install -e /absolute/path/to/KeepGPU` +1. Install: `pip install "git+https://github.com/Wangmerlyn/KeepGPU.git"` 2. Run: `keep-gpu --gpu-ids 0 --vram 1GiB --interval 60 --busy-threshold 25` 3. Verify: check CLI logs for keep loop activity; stop with `Ctrl+C` when done. From 65b8d7cddce26e276765d08f531c1d1320caed76 Mon Sep 17 00:00:00 2001 From: Wang Siyuan Date: Wed, 4 Mar 2026 12:56:19 +0800 Subject: [PATCH 5/7] docs(skills): prefer tmux for remote sessions --- skills/keepgpu-repo-workflow/SKILL.md | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-) diff --git a/skills/keepgpu-repo-workflow/SKILL.md b/skills/keepgpu-repo-workflow/SKILL.md index 76c4f0b..e5a0e92 100644 --- a/skills/keepgpu-repo-workflow/SKILL.md +++ b/skills/keepgpu-repo-workflow/SKILL.md @@ -116,10 +116,21 @@ All visible GPUs with lighter load: keep-gpu --vram 512MB --interval 180 ``` -Background execution on remote shell: +Remote sessions (preferred: `tmux` for visibility and control): + +```bash +tmux new -s keepgpu +keep-gpu --gpu-ids 0 --vram 1GiB --interval 300 +# Detach with Ctrl+b then d; reattach with: tmux attach -t keepgpu +``` + +Fallback when `tmux` is unavailable: ```bash nohup keep-gpu --gpu-ids 0 --vram 1GiB --interval 300 > keepgpu.log 2>&1 & +echo $! > keepgpu.pid +# Monitor: tail -f keepgpu.log +# Stop: kill "$(cat keepgpu.pid)" ``` ## Troubleshooting From e9eef73b8a06679e9945bf9aa0dd976f7ae3e340 Mon Sep 17 00:00:00 2001 From: Wang Siyuan Date: Wed, 4 Mar 2026 12:57:25 +0800 Subject: [PATCH 6/7] docs(skills): drop response template section --- .../skillcheck-free-report.md | 1 - skills/keepgpu-repo-workflow/SKILL.md | 20 ------------------- 2 files changed, 21 deletions(-) diff --git a/docs/skills/keepgpu-repo-workflow/skillcheck-free-report.md b/docs/skills/keepgpu-repo-workflow/skillcheck-free-report.md index 39d67d2..75ca4ad 100644 --- a/docs/skills/keepgpu-repo-workflow/skillcheck-free-report.md +++ b/docs/skills/keepgpu-repo-workflow/skillcheck-free-report.md @@ -12,7 +12,6 @@ Validation method: applied the SkillCheck Free rule set from `skill-check/SKILL. - Description includes activation triggers (8.3) - Description includes negative triggers (8.7) -- Skill specifies output format with examples (8.4) - Skill includes example section (8.1) - Skill documents error handling or limitations (8.2) - Skill uses structured instructions (8.5) diff --git a/skills/keepgpu-repo-workflow/SKILL.md b/skills/keepgpu-repo-workflow/SKILL.md index e5a0e92..ca9b97d 100644 --- a/skills/keepgpu-repo-workflow/SKILL.md +++ b/skills/keepgpu-repo-workflow/SKILL.md @@ -140,26 +140,6 @@ echo $! > keepgpu.pid - No utilization telemetry: ensure `nvidia-ml-py` works and `nvidia-smi` is available. - No GPUs detected: verify drivers, CUDA/ROCm runtime, and `torch.cuda.device_count()`. -## Output format - -Return guidance in this structure: - -```markdown -## KeepGPU CLI Plan - -### Install -- - -### Run Command -- - -### Verify -- - -### Stop -- Press Ctrl+C (or kill background PID) to release VRAM. -``` - ## Example User request: "Install KeepGPU from GitHub and keep GPU 0 alive while I preprocess." From a49a618cc470174a133eb72c212c12fb91f443a4 Mon Sep 17 00:00:00 2001 From: Wang Siyuan Date: Wed, 4 Mar 2026 13:08:19 +0800 Subject: [PATCH 7/7] docs(skills): rename skill to gpu-keepalive-with-keepgpu --- .../skillcheck-free-report.md | 4 ++-- .../SKILL.md | 2 +- .../agents/openai.yaml | 0 3 files changed, 3 insertions(+), 3 deletions(-) rename docs/skills/{keepgpu-repo-workflow => gpu-keepalive-with-keepgpu}/skillcheck-free-report.md (85%) rename skills/{keepgpu-repo-workflow => gpu-keepalive-with-keepgpu}/SKILL.md (99%) rename skills/{keepgpu-repo-workflow => gpu-keepalive-with-keepgpu}/agents/openai.yaml (100%) diff --git a/docs/skills/keepgpu-repo-workflow/skillcheck-free-report.md b/docs/skills/gpu-keepalive-with-keepgpu/skillcheck-free-report.md similarity index 85% rename from docs/skills/keepgpu-repo-workflow/skillcheck-free-report.md rename to docs/skills/gpu-keepalive-with-keepgpu/skillcheck-free-report.md index 75ca4ad..d17cc13 100644 --- a/docs/skills/keepgpu-repo-workflow/skillcheck-free-report.md +++ b/docs/skills/gpu-keepalive-with-keepgpu/skillcheck-free-report.md @@ -1,4 +1,4 @@ -# SkillCheck Free Report: keepgpu-repo-workflow +# SkillCheck Free Report: gpu-keepalive-with-keepgpu Validation method: applied the SkillCheck Free rule set from `skill-check/SKILL.md` in after updating the skill to focus on KeepGPU CLI usage and installation instructions. @@ -19,4 +19,4 @@ Validation method: applied the SkillCheck Free rule set from `skill-check/SKILL. ## Checked files -- `skills/keepgpu-repo-workflow/SKILL.md` +- `skills/gpu-keepalive-with-keepgpu/SKILL.md` diff --git a/skills/keepgpu-repo-workflow/SKILL.md b/skills/gpu-keepalive-with-keepgpu/SKILL.md similarity index 99% rename from skills/keepgpu-repo-workflow/SKILL.md rename to skills/gpu-keepalive-with-keepgpu/SKILL.md index ca9b97d..cbd9544 100644 --- a/skills/keepgpu-repo-workflow/SKILL.md +++ b/skills/gpu-keepalive-with-keepgpu/SKILL.md @@ -1,5 +1,5 @@ --- -name: keepgpu-repo-workflow +name: gpu-keepalive-with-keepgpu description: Install and operate the KeepGPU CLI to keep reserved GPUs active during data prep, debugging, and orchestration downtime. Use when users ask for keep-gpu command construction, tuning (--vram, --interval, --busy-threshold), installation from this repository, or runtime troubleshooting of keep-gpu sessions; do not use for repository development, code refactoring, or unrelated Python tooling. --- diff --git a/skills/keepgpu-repo-workflow/agents/openai.yaml b/skills/gpu-keepalive-with-keepgpu/agents/openai.yaml similarity index 100% rename from skills/keepgpu-repo-workflow/agents/openai.yaml rename to skills/gpu-keepalive-with-keepgpu/agents/openai.yaml