Show Me Your Work Decision Log
cursor/plugins
Keeps a TSV decision log for long or unattended agent runs, one row per decision with what, why, evidence and result, so a reviewer can check the work later.
Design the human approval surface for an agent system — which actions gate, how approvals batch without becoming rubber stamps, and what the audit trail must hold.
Install Human In The Loop Design
$ npx skills add mohitagw15856/pm-claude-skills --skill human-in-the-loop-design -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install mohitagw15856/pm-claude-skills human-in-the-loop-design --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/mohitagw15856/pm-claude-skills.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/human-in-the-loop-design .claude/skills/human-in-the-loop-design && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "human-in-the-loop-design" agent skill from https://github.com/mohitagw15856/pm-claude-skills/tree/main/skills/human-in-the-loop-design into .claude/skills/human-in-the-loop-design/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "human-in-the-loop-design", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/mohitagw15856/pm-claude-skills/tree/main/skills/human-in-the-loop-designType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add mohitagw15856/pm-claude-skills --skill human-in-the-loop-design -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install mohitagw15856/pm-claude-skills human-in-the-loop-design --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/mohitagw15856/pm-claude-skills.git skills-src && mkdir -p .agents/skills && cp -r skills-src/skills/human-in-the-loop-design .agents/skills/human-in-the-loop-design && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "human-in-the-loop-design" agent skill from https://github.com/mohitagw15856/pm-claude-skills/tree/main/skills/human-in-the-loop-design into .agents/skills/human-in-the-loop-design/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "human-in-the-loop-design", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add mohitagw15856/pm-claude-skills --skill human-in-the-loop-design -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install mohitagw15856/pm-claude-skills human-in-the-loop-design --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/mohitagw15856/pm-claude-skills.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/skills/human-in-the-loop-design .cursor/skills/human-in-the-loop-design && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "human-in-the-loop-design" agent skill from https://github.com/mohitagw15856/pm-claude-skills/tree/main/skills/human-in-the-loop-design into .cursor/skills/human-in-the-loop-design/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "human-in-the-loop-design", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/mohitagw15856/pm-claude-skills.git --path skills/human-in-the-loop-design--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add mohitagw15856/pm-claude-skills --skill human-in-the-loop-design -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install mohitagw15856/pm-claude-skills human-in-the-loop-design --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/mohitagw15856/pm-claude-skills.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/skills/human-in-the-loop-design .gemini/skills/human-in-the-loop-design && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "human-in-the-loop-design" agent skill from https://github.com/mohitagw15856/pm-claude-skills/tree/main/skills/human-in-the-loop-design into .gemini/skills/human-in-the-loop-design/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "human-in-the-loop-design", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install mohitagw15856/pm-claude-skills human-in-the-loop-designInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add mohitagw15856/pm-claude-skills --skill human-in-the-loop-design -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/mohitagw15856/pm-claude-skills.git skills-src && mkdir -p .github/skills && cp -r skills-src/skills/human-in-the-loop-design .github/skills/human-in-the-loop-design && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "human-in-the-loop-design" agent skill from https://github.com/mohitagw15856/pm-claude-skills/tree/main/skills/human-in-the-loop-design into .github/skills/human-in-the-loop-design/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "human-in-the-loop-design", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add mohitagw15856/pm-claude-skills --skill human-in-the-loop-design -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install mohitagw15856/pm-claude-skills human-in-the-loop-design --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/mohitagw15856/pm-claude-skills.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/skills/human-in-the-loop-design .opencode/skills/human-in-the-loop-design && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "human-in-the-loop-design" agent skill from https://github.com/mohitagw15856/pm-claude-skills/tree/main/skills/human-in-the-loop-design into .opencode/skills/human-in-the-loop-design/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "human-in-the-loop-design", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Facts
- Skill name
human-in-the-loop-design- GitHub stars
- 1.4k
- Token cost
- ~1.5k tokens
- SKILL.md length
- 757 words
- Files
- 1
- Skills in repo
- 1,348
- Repo updated
- First seen
- Licence
- MIT
At a glance
Design the human approval surface for an agent system — which actions gate, how approvals batch without becoming rubber stamps, and what the audit trail must hold.
- Works in 6 steps: Tier every action by consequence, not by… → Budget the approvals. Multiply… → Design the approval moment against… → …
- Asked to add human oversight to an agent
- SKILL.md covers What This Skill Produces, Required Inputs, Design Method and Output Format, plus 3 more sections
- Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md
What it does
Human In The Loop Design is an agent skill from mohitagw15856/pm-claude-skills. Design the human approval surface for an agent system — which actions gate, how approvals batch without becoming rubber stamps, and what the audit trail must hold. Use when asked to add human oversight to an agent, design approval workflows for AI actions, decide what an agent may do autonomously, or fix approval fatigue in an existing loop. Produces an action-tier policy, approval UX spec, escalation rules, and audit-trail requirements. For specifying the whole agent use agent-spec; for the per-skill execution…
Its SKILL.md is about 1.5k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.
It sits in Agent Workflows, covering Human-in-the-loop approvals. The repository describes itself as: 1255 professional Agent Skills for Claude, ChatGPT, Gemini, Cursor & Codex — PRDs, postmortems, leases, medical bills, layoffs, go-bags, new countries. Plain markdown, MIT, in… The licence is MIT.
When your agent uses it
- Asked to add human oversight to an agent
- Design approval workflows for AI actions
- Decide what an agent may do autonomously
- Fix approval fatigue in an existing loop
Example prompts
- “/human-in-the-loop-design”
Workflow steps
6 steps, taken from the first numbered list in SKILL.md.
- Tier every action by consequence, not by feel. Two axes decide the tier: reversibility (undo in one step ↔ irreversible) and reach…
- Budget the approvals. Multiply approve-tier actions by daily volume. If the number exceeds ~10-15 meaningful decisions per approver per…
- Design the approval moment against rubber-stamping.
- Write the escalation rules. Approver silent for [X]: action expires safely (never auto-proceeds). Approver rejects: agent gets the reason…
- Spec the audit trail. Per gated action: what the agent proposed (verbatim), the evidence it showed, who decided, what shipped (diff vs…
- Plan the tier reviews. Tiers loosen with evidence, not with comfort: an action moves down a tier after [N] consecutive approvals with zero…
What it can do on your machine
Read from SKILL.md and the folder at commit 1cbf1f0. It shows what the files ask for, not the result of running them.
Tool permissions
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Runs code
No scripts in the folder and no shell commands in SKILL.md.
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Context cost
Human In The Loop Design loads about 1.5k tokens when it runs. Until then it costs about 149 tokens; SKILL.md has 757 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
Safety
Auto-check passed
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
SKILL.md
The full file from mohitagw15856/pm-claude-skills at commit 1cbf1f0, republished under its MIT licence (© mohitagw15856). 757 words, ~1,532 tokens.
Download SKILL.mdSave it as
.claude/skills/human-in-the-loop-design/SKILL.md (or your agent's skills folder).- name
- human-in-the-loop-design
- description
- Design the human approval surface for an agent system — which actions gate, how approvals batch without becoming rubber stamps, and what the audit trail must hold. Use when asked to add human oversight to an agent, design approval workflows for AI actions, decide what an agent may do autonomously, or fix approval fatigue in an existing loop. Produces an action-tier policy, approval UX spec, escalation rules, and audit-trail requirements. For specifying the whole agent use agent-spec; for the per-skill execution gates see the Execution-block pattern in SKILLSPEC §5.
Human-in-the-Loop Design Skill
The failure mode of agent oversight isn't too little review — it's review that decays into a rubber stamp. Forty approval prompts a day trains the human to click yes; then the one that mattered goes through with the rest. This skill designs the loop so human attention lands exactly where it changes the outcome, and nowhere else.
What This Skill Produces
- An action-tier policy: every agent action classified auto / notify / approve / forbidden
- An approval UX spec: what the human sees, batching rules, and the anti-rubber-stamp mechanics
- Escalation & fallback rules: timeouts, absent approvers, disagreement
- Audit-trail requirements: what gets recorded so any decision is reconstructable
Required Inputs
Ask for (if not already provided):
- The agent and its action inventory — everything it can do (from its tool list, not its marketing)
- Blast radius per action: reversible? outward-facing? money/data/permissions involved?
- Volume estimates: how many times per day each action fires (approval load is a design constraint, not an afterthought)
- Who approves — role, how many people, what else competes for their attention
Design Method
- Tier every action by consequence, not by feel. Two axes decide the tier: reversibility (undo in one step ↔ irreversible) and reach (internal draft ↔ external/financial/permanent). Then:
- Auto — reversible + internal (drafts, reads, internal scratch writes). Log only.
- Notify — reversible + modest reach (filed a ticket, updated a record). Do it, tell the human, easy undo.
- Approve — hard to reverse OR outward-facing (send, publish, pay, delete, grant). Blocks until a human decides.
- Forbidden — irreversible + high reach where the org has decided no automation belongs (auth changes, legal commitments). Not gated — absent from the toolset.
- Budget the approvals. Multiply approve-tier actions by daily volume. If the number exceeds ~10-15 meaningful decisions per approver per day, the design is broken before launch: move volume down-tier by adding reversibility (drafts, holds, delayed sends) rather than by lowering the bar.
- Design the approval moment against rubber-stamping.
- Show the decision, not the transcript: what will happen, to whom, why the agent believes it's right, and what's unusual about this one.
- Surface anomaly, hide routine: same-as-last-50 approvals batch into one digest; the outlier renders differently and alone.
- Require typed engagement for the highest stakes (type the amount, name the recipient) — friction proportional to consequence.
- Track approval latency and edit rate per approver: 100% instant approvals is a broken loop, not a good agent — say so in the metrics section.
- Write the escalation rules. Approver silent for [X]: action expires safely (never auto-proceeds). Approver rejects: agent gets the reason as context, may revise once, then stops. Two approvers disagree: named tiebreaker. After-hours urgent: the on-call path, or an honest "waits until morning".
- Spec the audit trail. Per gated action: what the agent proposed (verbatim), the evidence it showed, who decided, what shipped (diff vs proposal), timestamps. The reconstruction test: six months later, "why did this go out?" is answerable from the trail alone.
- Plan the tier reviews. Tiers loosen with evidence, not with comfort: an action moves down a tier after [N] consecutive approvals with zero edits and a human review of a sample. Tightening is immediate on any incident.
Show full SKILL.md (240 more words)Show less
Output Format
HITL Design: [agent system]
Action-tier policy
| Action | Reversibility | Reach | Tier | Volume/day | Notes |
|---|
Approval load: [decisions/day/approver at launch — and the redesign applied if it exceeded budget]
The approval moment: [what renders; batching rules; anomaly surfacing; typed-engagement thresholds]
Escalation: [timeout → outcome · rejection → protocol · disagreement → tiebreaker · after-hours → path]
Audit trail: [fields recorded; retention; who can query]
Tier evolution: [down-tier evidence bar · instant up-tier triggers · review cadence]
Health metrics: [approval latency, edit rate, override rate — with the "100% instant approvals means the loop is dead" alarm]
Quality Checks
- Every action in the agent's toolset appears in the tier table — none defaulted silently
- The approval budget is computed, and the launch design fits inside it
- Timeout behaviour is safe-by-default (expire, never auto-proceed)
- The forbidden tier removes capabilities from the toolset rather than gating them
- Health metrics detect rubber-stamping, not just agent errors
Anti-Patterns
- Do not gate everything — undifferentiated approval load is how the important one slips through
- Do not show raw transcripts as the approval artifact — humans approve decisions, not logs
- Do not let unanswered approvals auto-proceed on timeout "to keep things moving"
- Do not loosen tiers on gut feel — the down-tier bar is written evidence, the up-tier trigger is any incident
- Do not measure only agent mistakes — an approver who edits nothing for a month is the riskier signal
Example Trigger Phrases
- "Design approval workflows for AI actions."
- "Decide what an agent may do autonomously."
- "Fix approval fatigue in an existing loop."
© mohitagw15856, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Files
Just SKILL.md in skills/human-in-the-loop-design of mohitagw15856/pm-claude-skills.
Open the folder on GitHubat commit 1cbf1f0
Compare with similar skills
Human In The Loop Design next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Human In The Loop Design this skillmohitagw15856/pm-claude-skills | 1.4k | — | ~1.5k | Automated safety check: Pass | MIT | |
| Show Me Your Work Decision Logcursor/plugins | 11k | 8 repos | ~1.6k | Automated safety check: Pass | None | |
| Darwin Skill Optimizeralchaincyf/darwin-skill | 6.2k | 1 repos | ~4.7k | Automated safety check: Pass | MIT | |
| Loop Constraints Enforcercobusgreyling/loop-engineering | 11k | 1 repos | ~475 | Automated safety check: Notes | MIT | |
| Ask User QuestionMemTensor/MemOS | 12k | — | ~1k | Automated safety check: Pass | Apache-2.0 | |
| PUA High-Agency Governancetanweai/pua | 20k | — | ~502 | Automated safety check: Pass | MIT |
Similar skills
- Official
Darwin Skill Optimizer
alchaincyf/darwin-skill
Scores SKILL.md files on a nine-dimension rubric, then improves them in a keep-or-revert loop with independent judge agents, test prompts, git history and human checkpoints.
Loop Constraints Enforcer
cobusgreyling/loop-engineering
Loads a project's loop-constraints.md before any other action and blocks pushes, edits or merges that violate the rules it defines.
Ask User Question
MemTensor/MemOS
Shows a question as a modal in the interface to clarify a task, collect a preference or get approval, since the user cannot see terminal output.
PUA High-Agency Governance
tanweai/pua
Pushes an agent to keep verifying and changing approach after repeated failures, using a diagnosis line, evidence-based completion and confirmation before risky edits.
Agentmemory Forget
rohitg00/agentmemory
Deletes chosen memories from agentmemory only after showing the matches and getting an explicit yes, for privacy requests and cleanup of outdated notes.
More from mohitagw15856/pm-claude-skills
All 1,348 skills in this repoCar Tco
mohitagw15856/pm-claude-skills
Compare the total cost of car ownership across buy-new, buy-used, lease, and keep-your-current-car — depreciation, insurance, maintenance ramp, and fuel over a real horizon, not just the monthly…
Cs Health Scorecard
mohitagw15856/pm-claude-skills
Build a customer health scorecard for a specific account. An agent skill from mohitagw15856/pm-claude-skills.
Exit Waterfall
mohitagw15856/pm-claude-skills
Compute who gets what at each exit price from a cap table — liquidation preferences, conversion points, and where the founders' share collapses.
Feature Prioritisation
mohitagw15856/pm-claude-skills
Apply prioritisation frameworks (RICE, MoSCoW, Kano, ICE, Opportunity Scoring) to rank features and backlog items.
Fire Number
mohitagw15856/pm-claude-skills
Compute a financial-independence (FIRE) target and years-to-reach with every assumption labeled as an assumption — plus a sensitivity table instead of a single false-precision answer.
Freelance Rate
mohitagw15856/pm-claude-skills
Derive a freelance day/hourly rate backwards from target income, honest billable utilization, overhead, and the self-employment tax premium — the arithmetic that proves a rate is not salary÷2000.
Related
Categories
Questions about Human In The Loop Design
What does Human In The Loop Design do?
Design the human approval surface for an agent system — which actions gate, how approvals batch without becoming rubber stamps, and what the audit trail must hold. Human In The Loop Design is an agent skill from mohitagw15856/pm-claude-skills. Design the human approval surface for an agent system — which actions gate, how approvals batch without becoming rubber stamps, and what the audit trail must hold.
When should I use Human In The Loop Design?
Human In The Loop Design fits situations like: asked to add human oversight to an agent; design approval workflows for AI actions; decide what an agent may do autonomously; fix approval fatigue in an existing loop.
How do I install Human In The Loop Design in Claude Code?
Run `npx skills add mohitagw15856/pm-claude-skills --skill human-in-the-loop-design -a claude-code`. Or copy the skill folder (skills/human-in-the-loop-design in mohitagw15856/pm-claude-skills) into .claude/skills/human-in-the-loop-design in your project. Claude Code loads it when a task matches its description.
How do I install Human In The Loop Design in Codex?
Run `npx skills add mohitagw15856/pm-claude-skills --skill human-in-the-loop-design -a codex`. Or copy the skill folder (skills/human-in-the-loop-design in mohitagw15856/pm-claude-skills) into .agents/skills/human-in-the-loop-design in your project. Codex loads it when a task matches its description.
Can I use Human In The Loop Design in Cursor, Gemini CLI or GitHub Copilot?
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add mohitagw15856/pm-claude-skills --skill human-in-the-loop-design -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/human-in-the-loop-design, .gemini/skills/human-in-the-loop-design, .github/skills/human-in-the-loop-design and .opencode/skills/human-in-the-loop-design in your project.
What does Human In The Loop Design need to run?
SKILL.md names no scripts, command-line tools or credentials: Human In The Loop Design is instructions for the agent only.
Does Human In The Loop Design access the network?
SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.
Is Human In The Loop Design safe to install?
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.
What licence does Human In The Loop Design use?
Human In The Loop Design is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
How many tokens does Human In The Loop Design use?
About 1.5k tokens (SKILL.md is roughly 6.1k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.
What are the alternatives to Human In The Loop Design?
Skills that share tags, products or a category with Human In The Loop Design: Show Me Your Work Decision Log (cursor/plugins, 11k stars), Darwin Skill Optimizer (alchaincyf/darwin-skill, 6.2k stars), Loop Constraints Enforcer (cobusgreyling/loop-engineering, 11k stars) and Ask User Question (MemTensor/MemOS, 12k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Human In The Loop Design?
mohitagw15856 (a GitHub user) maintains it in mohitagw15856/pm-claude-skills, which has 1,434 GitHub stars. The repository holds 1,348 skills in this directory. The repository was last updated on October 9, 2026.
Source: mohitagw15856/pm-claude-skills on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.