Agent Skill
Hermes Ephemeral Delegation skill
hermes-ephemeral-delegationGentleman-Programming/gentle-aiinternal/assets/skills/hermes-ephemeral-delegation/SKILL.md
Trigger: a mapping need, parallel units, context backstop, high-risk verify, fresh review, or multi-step debug. Delegate via delegate_task for a named reason.
always on
43 tokname + description
once used
903 tokSKILL.md body
bundled files
11 KB
repo stars
7.9kGitHub
Install the Hermes Ephemeral Delegation skill
Skills CLI (Claude Code, Codex, Cursor and others)
npx skills add https://github.com/Gentleman-Programming/gentle-ai --skill hermes-ephemeral-delegation
Claude Code, by hand (all projects)
git clone --depth 1 https://github.com/Gentleman-Programming/gentle-ai.git /tmp/gentle-ai mkdir -p ~/.claude/skills cp -R /tmp/gentle-ai/internal/assets/skills/hermes-ephemeral-delegation ~/.claude/skills/hermes-ephemeral-delegation
For one project only, copy the folder into .claude/skills/ in that project instead. The agent picks it up when a task matches its description.
What it costs your context
Agent Skills load in three steps. Tokens counted with the o200k tokenizer; your model's count will differ a little.
| When | What loads | Cost |
|---|---|---|
| Every session | Name and description | 43 tok |
| When the agent uses it | The SKILL.md instructions | 903 tok |
| Only if opened | 1 bundled files (1 KB) | varies |
Until it is used, hermes-ephemeral-delegation costs about 32× less than the median MCP server we measured (1,362 tokens on every turn). Installing all 26 skills from Gentleman-Programming/gentle-ai would load 1,022 tokens of descriptions. Skills vs MCP study
Checks
Our checks found nothing to flag: it follows the Agent Skills specification's naming and size guidelines and doesn't pipe downloads into a shell.
Checked 2026-10-22 · method
What's inside
The headings of its SKILL.md. Read the full file on GitHub.
- Activation Contract
- Hard Rules
- Decision Gates
- Execution Steps
- Output Contract
- References
More skills from Gentleman-Programming/gentle-ai
- branch-prGentleman-Programming/gentle-ai
Create Gentle AI pull requests with issue-first checks. Trigger: creating, opening, or preparing PRs for review.
7.9k26 tok idle · 2,213 in use - chained-prGentleman-Programming/gentle-ai
Trigger: PRs over 400 lines, stacked PRs, review slices. Split oversized changes into chained PRs that protect review focus.
7.9k32 tok idle · 642 in use - cognitive-doc-designGentleman-Programming/gentle-ai
Design docs that reduce cognitive load. Trigger: writing guides, READMEs, RFCs, onboarding, architecture, or review-facing docs.
7.9k33 tok idle · 459 in use - comment-writerGentleman-Programming/gentle-ai
Write warm, direct collaboration comments. Trigger: PR feedback, issue replies, reviews, Slack messages, or GitHub comments.
7.9k29 tok idle · 467 in use - gentle-ai-benchGentleman-Programming/gentle-ai
Trigger: bench, journey, journeys, driven mode, gentle-ai-bench, journey corpus, j-numbers, bench axis. Author and verify gentle-ai bench journeys; go test ./bench never proves driven execution.
7.9k49 tok idle · 624 in use - gentle-ai-branch-prGentleman-Programming/gentle-ai
Create Gentle AI pull requests with issue-first checks. Trigger: creating, opening, or preparing PRs for review.
7.9k30 tok idle · 2,994 in use - gentle-ai-chained-prGentleman-Programming/gentle-ai
Trigger: PRs over 400 lines, stacked PRs, review slices. Split oversized changes into chained PRs that protect review focus.
7.9k35 tok idle · 503 in use - gentle-ai-collab-perfectGentleman-Programming/gentle-ai
Trigger: contributing to Gentleman-Programming/gentle-ai as an external collaborator. Strict issue-first workflow, honest PR bodies, contributor-vs-maintainer scope, chained-PR strategy, verification protocol, docstring coverage. Load whenever the active repo is Gentleman-Programming/gentle-ai and any part of the contribution flow is in scope: opening an issue, drafting or editing a PR body, splitting a change into chained/stacked PRs, or auditing a PR before requesting review.
7.9k106 tok idle · 4,820 in use