Dynamic Workflow Mode Harnesses
Stop improvising agent workflows with Muse: design task-local harnesses with objective, inputs, outputs, eval gates, and handoff artifacts, then extract shared skills only when the pattern repeats.
- What
- Stop improvising agent workflows with Muse: design task-local harnesses with objective, inputs, outputs, eval gates, and handoff artifacts, then extract shared skills only when the pattern repeats.
- Cost
- Free
- Needs
- none. No account, no API key, no payment — this skill is a single guidance file.
- Level
- Beginner
- Install
- Copy the installer prompt below into your Muse — your agent does the rest.
Curated by Skill Harbor: a discipline for Claude dynamic workflow mode that turns ad-hoc agent improvisation into an observable system. The core contract: a task-local harness is only worth building when it is cheaper and safer than manually driving the same steps, and it must declare its Objective (what it owns and explicitly does not own), Inputs (files, URLs, prompts, data sources, credentials policy, constraints), Outputs (commits, reports, screenshots, status files, control pane snapshots), Eval (at least one pass/fail check tied to the task, never just "it ran"), and Handoff (a short artifact telling the next operator what happened, what is blocked, and how to resume). The decision tree keeps you honest: one-shot task stays inline, repeated task with changing inputs gets a task-local harness in a temp or project-local area, repeated task across teammates or repos gets extracted into a shared skill, tasks with external state or approvals get control pane visibility first, and safety-risky tasks get an eval gate plus a human merge gate before autonomous execution. Control pane checkpoints (Plan, Queue, Run, Gate, Handoff) make the workflow team-usable across sessions, and eval gates are chosen per work type: focused tests for code, browser smoke for UI, fixture transcripts for agent workflows, claim checklists for research, dry runs for integrations. Shared skill extraction is gated too: promote only when at least two signals hold (the workflow repeats across sessions or teams, failures repeat because operators skip a gate, the input/output contract is stable). Anti-patterns are named: scripts that hide decision logic from the operator, treating dynamic mode as permission to skip tests, one-off docs where a shared skill is the real product, multiple agents without ownership or merge policy, raw private research leaking into public docs. By @affaan-m, listed here with credit to its creator. From the affaan-m/ECC repository (MIT). Honest caveats: never claim a workflow is reusable until the eval can be rerun by another teammate; a harness without an eval gate is just automation with extra steps. The template is a starting structure, adapt the loop to your task. Skill Harbor never reviews the code, review it yourself before use.
Version:
Quick answer
Dynamic Workflow Mode Harnesses is a free Muse skill and discipline that turns ad-hoc agent improvisation into an observable system. Its core contract is that a task-local harness is only worth building when cheaper and safer than manual driving, and it must declare objective, inputs, outputs, eval gates and handoff artifacts.
Install
Copy the install package below, then paste it into Muse🧪 Install ran in our test sandbox · Nov 11, 2026📌 Pinned version: commit 4eb71d9Step 0 — before installing anything: read this install prompt, summarize in a few lines what this build does and what installing it involves, then ask me to confirm before you proceed. Prerequisites: none. No account, no API key, no payment — this skill is a single guidance file. Install "Dynamic Workflow Mode Harnesses" for me. It shows my agent how to design task-local harnesses, eval gates and reusable skill extraction for dynamic workflow mode and other adaptive agent harnesses. Once it is installed, I use it by asking my agent to build a task-local harness or to extract a shared skill from repeated work. Repository: https://raw.githubusercontent.com/affaan-m/ECC/4eb71d92a39cab44ad40ac9d8a6a5ccb4029d6c2/skills/dynamic-workflow-mode/SKILL.md Pinned version: install the code at commit 4eb71d92a39cab44ad40ac9d8a6a5ccb4029d6c2 exactly (Skill Harbor tested this version). Do not substitute a newer version. If this commit is unreachable, stop and tell me. 1. Fetch the SKILL.md file from the repository path into a temporary folder and look inside. It is the only file this skill needs. 2. Integrity check: Skill Harbor reviewed this listing on 2026-10-10 and recorded the fingerprint below. I want you to compute the SHA-256 of the downloaded file and compare: - SKILL.md: SHA-256 bdf7902e4e213c6f2e7bdc417757004d118e8e36267239b707670e2875ebf7bf If a fingerprint does not match, I want you to stop before installing and tell me the files changed since Skill Harbor's review, so the new version is unverified. Show me your safety findings and give me three choices: (a) wait for Skill Harbor to re-review the listing (recommended), (b) analyze the new files in detail first and then decide, (c) install anyway knowing the risk. Then wait for my answer before doing anything else. 3. Safety check: I want to see your findings explicitly, in short separate items: - every network address the code contacts (expected: none, this file is guidance text only) - every shell command it runs (expected: none) - every place credentials, API keys or personal data are read, stored or sent (expected: none) - any obfuscated or encoded code (expected: none, plain markdown guidance) If anything looks suspicious, stop and explain it to me before continuing. 4. Confirmation: before installing, give me a short summary of what the skill does plus your safety findings, and wait for my explicit "yes, install". 5. Install it as a skill: copy SKILL.md into the agent's skills directory, in a folder named "dynamic-workflow-mode". 6. Verify with no network calls: frontmatter valid, skill name is dynamic-workflow-mode, and the file is in place. 7. Report what was installed, where, and what I still need to do myself. Nothing else is required: to use the skill afterwards, I describe a workflow to my agent and ask it to design a task-local harness for it, or to extract a reusable skill from repeated work. GitHub is optional: if I have a GitHub account or the gh CLI, you may use it; otherwise public access is fine. Never require it unless it's in the prerequisites above. Rules: don't touch anything outside the temp folder and the install target. If anything looks off, stop and ask me.
Saved to your recent installs. Find it anytime on /connect.
🔒 Security, provenance and what this touches
The install prompt below already includes the vetting steps: your agent follows the community checklist before installing anything with executable code. Want more?
Verification certificate
Dated facts about what was checked for this listing, and how. This is not a safety certification.
- Pinned version
- 4eb71d9 · 2026-10-13
- Sandbox install test
- Install ran in our test sandbox · Nov 11, 2026
- Creator identity
- Not verified
- Fingerprint watch
- Last checked · Nov 21, 2026 : no change detected in the files linked by this install prompt
- Install prompt analysis
- 2026-10-13 : No commands, network calls, packages, file writes or API keys detected.
Skill Harbor never reviews the code of a listed skill. These are dated facts about the install prompt and our install test, not a safety certification. Review a skill yourself before you install it. How we verify
Questions
How do I install a skill?
Every product page includes a copy-paste install prompt. Paste it into your Muse and it sets the skill up for you — no manual configuration.
Where does my money go?
Straight to the seller. Skill Harbor never processes payments: checkout happens on the seller’s own page, usually Stripe.
What does the ✓ next to a creator’s name mean?
It means we confirmed the identity of the person behind the listing. It says nothing about the code itself — always check a skill before installing it.