Not in the top 1,000 by GitHub stars
ray-amjad/skills
Skills for coding agents. Straight from my own Claude Code and Codex setup.
History
Review weekly changes to SKILL.md before updating. Dates mark the weeks in which we observed each version; checks apply to the recorded commits.
View weekly changes
skills/audit-prove-it/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 0f7e404357e8
No earlier version was observed for comparison.
skills/create-prove-it/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 0f7e404357e8
No earlier version was observed for comparison.
skills/maintain-prove-it/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 0f7e404357e8
No earlier version was observed for comparison.
View on GitHubAdd to my skills
Verified commit 0f7e404357e8f5676b5e4a6fcac3560395be8795
git clone --filter=blob:none --no-checkout https://github.com/ray-amjad/skills.git .hotskills/ray-amjad-skills && git -C .hotskills/ray-amjad-skills checkout --detach 0f7e404357e8f5676b5e4a6fcac3560395be8795
# Copy each selected skill directory from .hotskills/ray-amjad-skills into .claude/skills/<skill-name>/59GitHub stars
Not availablestars in 7 days
3skills in this repository
18commits in 30 days
Heat
Insufficient historyNot enough history yet to measure heat. It appears after a few days of data.
Heat never changes the order by stars, and sponsorship never touches it. How heat works
Stars over time
30-day change not yet available: not enough history.
Individual skills in this repository
- audit-prove-it
skills/audit-prove-it - create-prove-it
skills/create-prove-it - maintain-prove-it
skills/maintain-prove-it
How it works
audit-prove-it
When your agent uses it
Try to fool a repo's verification skill, either kind made by /create-prove-it (the permanent prove-<app> or a temporary prove-<migration>). Plants one deliberate bug at a time in the files the current change touched (or along a flow the user names), and for a migration only in the new system, never the old one it is compared with. A fresh agent then checks that change or flow with the skill and its answer keys, without knowing what was planted. Records whether it caught the bug, reverts exactly, and reports every miss with a suggested fix to the skill and to the flow's answer key. Use for /audit-prove-it, "can the verifier be fooled", "test the verify skill", "is our verification any good", "mutation test the verifier", or after creating or heavily editing a prove-* skill.
What it loads into context
Frontmatter
183 tokensAlways available to the agent
SKILL.md
2,066 tokensWhen the skill is activated
Reference documents
16,019 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Safety first
- 1. Find the skill and baseline it
- 2. Plan the bugs
- 3. Plant, check blind, revert
- 4. Report
- 5. Offer to close the gaps
Files
../../.claude-plugin/marketplace.jsonasset431 bytes../../.claude-plugin/plugin.jsonasset617 bytesSKILL.mdcore8,725 bytesagents/openai.yamlasset124 bytesreferences/mutation-catalogue.mddoc2,566 bytes../create-prove-it/SKILL.mddoc23,454 bytes../create-prove-it/agents/openai.yamlasset134 bytes../create-prove-it/references/methods.mddoc8,931 bytes../create-prove-it/references/skill-template.mddoc18,575 bytes../create-prove-it/references/unreachable-surfaces.mddoc4,176 bytes../maintain-prove-it/SKILL.mddoc9,946 bytes../maintain-prove-it/agents/openai.yamlasset147 bytes
Read SKILL.md
# Audit Prove It
A verification skill that says PROVED on a broken app is worse than none: it gives false confidence. The only way to know it works is to break the app on purpose and see if it notices. This skill does that, safely, and reports every bug that slipped past.
## Safety first
You are about to edit product code on purpose. These rules keep that from hurting anyone:
- **Only in a git repo, only on local files.** Never against a shared, staging or production instance.
- **Back up before you plant.** The files a change touched usually have uncommitted work in them, and `git checkout` would throw it away. Before planting, copy the file byte for byte to `.prove-it/audits/<UTC timestamp>/originals/<path>`. Revert by copying it back. This is what makes the revert exact.
- **One planted bug at a time.** Plant, check, revert, confirm the revert, then the next.
- **Confirm every revert** with `cmp <backup> <file>` (exit 0) before moving on. If it doesn't match, stop everything and restore from the backup. Never use `git checkout` on a file that had uncommitted changes.
- **Never commit, push or stash a planted bug.** At the end, `git status` must look exactly as it did at the start, plus any receipts in `.prove-it/` and the report.
create-prove-it
When your agent uses it
Generate a project-local verification skill that an agent runs at the end of a change, before it says done. It asks first whether the skill is for the app or for a migration. For the app it writes or extends one permanent prove-<app> skill with a section per surface (web, API, MCP, CLI), and a new surface adds a section, not a new skill. For a migration (Python to Rust, Next.js to React Router) it writes a temporary prove-<migration> skill that runs old and new side by side, then hands its cases to prove-<app> as snapshots when the old system is deleted. It checks what the change touches (the new flows it adds and the existing flows it touched), each against an answer key the builder did not write, such as an old system, a snapshot, a test suite or the user's spec. Use for /create-prove-it, "set up verification", "make a verify skill for this repo", "I'm adding an MCP server" or any new surface, "verify this migration", or "the agent keeps saying done without testing".
What it loads into context
Frontmatter
236 tokensAlways available to the agent
SKILL.md
5,517 tokensWhen the skill is activated
Reference documents
7,621 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- What the change touches
- 0. The app, or a migration?
- 1. Map the surfaces
- 2. Check you can reach them
- 3. Find the answer keys
- 4. Brainstorm, then let the user pick
- 5. Generate the skill
- 6. Run it once for real
- 7. Hand over
Files
SKILL.mdcore23,454 bytesSKILL.mdcore23,454 bytesagents/openai.yamlasset134 bytesreferences/methods.mddoc8,931 bytesreferences/skill-template.mddoc18,575 bytesreferences/unreachable-surfaces.mddoc4,176 bytes
Read SKILL.md
# Create Prove It
You are about to give this repo a skill that makes every future agent prove its work in the running app, instead of saying "done" because the code looks right or the tests pass.
The output is a project-local skill, read cold by a future agent in the middle of some other task. Write it for that reader: exact commands, real selectors, no placeholders, nothing it has to work out.
The steps, in order:
0. Ask: the app, or a migration?
1. Map the surfaces
2. Check you can reach them
3. Find the answer keys
4. Brainstorm methods, then let the user pick
5. Generate the skillmaintain-prove-it
When your agent uses it
Improve a repo's generated verification skill from what its past runs learned. Works on both kinds made by /create-prove-it, the permanent prove-<app> and a temporary prove-<migration>. Reads the receipts in .prove-it/runs/ (failed and blocked steps, retries, the commands that finally worked, steps that failed for reasons unrelated to the change, how each change was sorted into checks) and any bug that slipped past prove, then fixes drifted steps, writes in the working commands, adds the check that would have caught a slipped bug and a section for any new surface. For a temporary skill it checks whether the migration is done and, if so, suggests the end step. Never rewrites an answer key (spec, snapshot, old system) to match the app. Proves every edit live and reports product bugs without touching product code. Use for /maintain-prove-it, "update the verify skill", "the verifier is out of date", "prove missed this bug", "prove keeps failing on unrelated steps", or after a big merge or migration milestone.
What it loads into context
Frontmatter
230 tokensAlways available to the agent
SKILL.md
2,297 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rules for the whole pass
- 1. Find the skill
- 2. Read the past runs
- 3. For a temporary skill, check whether the migration is done
- 4. Collect bugs that slipped past
- 5. Sort what you found
- 6. Edit and prove live
- 7. Report
Files
SKILL.mdcore9,946 bytesagents/openai.yamlasset147 bytes
Read SKILL.md
# Maintain Prove It
A verification skill goes stale the day the app changes: a route moves, a button gets renamed, a new surface ships with nothing in the skill that reaches it. A stale skill is worse than none, because the next agent will trust it. Every prove run leaves a receipt that shows where the skill struggled. This pass reads those receipts and folds what they learned back into the skill.
## Rules for the whole pass
- **Only edit the skill's own folder** (SKILL.md, scripts/). Never change product code here.
- **When the app and the skill disagree, work out which one is wrong.** The skill describing something the app no longer does is either *drift* (the app changed on purpose: fix the skill) or a *product bug* (the app broke: report it, don't rewrite the skill to match the bug). Read the commit that changed it. If you still can't tell, ask.
- **Answer keys are not the skill's steps.** The user's spec, the snapshots in `.prove-it/snapshots/` and the old system a flow is compared with say what right looks like. A drifted step can be fixed, but an answer key that disagrees with the app is reported to the user, never rewritten to match. Only the user changes one.
- **Every edit you make to the skill gets proved live before you finish.** An unproven edit is just more drift.
- **Old verdicts are leads, not evidence.** Earlier receipts, closed tickets and status docs tell you where to look. Only this pass's live runs count.
- **Never loosen a check to make a flaky step pass.** Record the failing run, then decide whether the app has a race or the check is ambiguous.
## 1. Find the skillSkills included
- audit-prove-it
- create-prove-it
- maintain-prove-it
Repository facts
- License
- MIT
- Created
- May 10, 2026
- Last pushed
- Nov 3, 2026
- Forks
- 7
- Open issues
- 0
- Project activity
- Active
Before you install
Network or broad shellAutomated signals, not a guarantee
Checked Nov 6, 2026
Not fully analyzed: some files were too long, unreadable or missing. This never lowers the level.
Show 1 note
skills/create-prove-it/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzed
- No broad tool permissions
- No dangerous shell commands
- 20 network referencesin 4 files: skills/create-prove-it/references/methods.md:90 skills/create-prove-it/references/skill-template.md:75 .claude-plugin/plugin.json:3 and 1 more
- No hidden or obfuscated text
- No instruction overrides
- No executable scripts
- No access to secrets
- No requests to vote in Hot
Show all 20 signals by file
.claude-plugin/marketplace.json
- Network accessline 1
{ "name": "ray-amjad", "owner": { "name": "Ray Amjad", "url": "https://www.agenticcoding.school" }, - Network accessline 2
"name": "ray-amjad", "owner": { "name": "Ray Amjad", "url": "https://www.agenticcoding.school" }, "description": "Ray Amjad's agent skills, as an installable plugin.", - Network accessline 3
"owner": { "name": "Ray Amjad", "url": "https://www.agenticcoding.school" },
.claude-plugin/plugin.json
- Network accessline 3
"version": "0.2.1", "description": "Ray Amjad's agent skills. Starts with Prove It: generate, maintain and audit a verification skill that makes your coding agent prove its work.", "author": { "name": "Ray Amjad", "url": "https://www.ag - Network accessline 4
"description": "Ray Amjad's agent skills. Starts with Prove It: generate, maintain and audit a verification skill that makes your coding agent prove its work.", "author": { "name": "Ray Amjad", "url": "https://www.agenticcoding.school" }, - Network accessline 5
"author": { "name": "Ray Amjad", "url": "https://www.agenticcoding.school" }, - Network accessline 6
"homepage": "https://www.agenticcoding.school", - Network accessline 7
"repository": "https://github.com/ray-amjad/skills",
skills/create-prove-it/references/methods.md
- Network accessline 90
- **Catches:** the page that jumps as it loads, the slow query a change added, the button with no label, the warning nobody reads. Things that work but are worse. - **Misses:** whether the feature does the right thing. Pair it with a drive, - Network accessline 90
- **Catches:** the page that jumps as it loads, the slow query a change added, the button with no label, the warning nobody reads. Things that work but are worse. - **Misses:** whether the feature does the right thing. Pair it with a drive, - Network accessline 91
- **Misses:** whether the feature does the right thing. Pair it with a drive, never use it alone. - **Cost:** low to medium. Pick a tool the agent can run every time (Chrome DevTools MCP, Lighthouse CLI, axe, `curl -w`), and measure more th - Network accessline 91
- **Misses:** whether the feature does the right thing. Pair it with a drive, never use it alone. - **Cost:** low to medium. Pick a tool the agent can run every time (Chrome DevTools MCP, Lighthouse CLI, axe, `curl -w`), and measure more th - Network accessline 92
- **Cost:** low to medium. Pick a tool the agent can run every time (Chrome DevTools MCP, Lighthouse CLI, axe, `curl -w`), and measure more than once when the number is noisy. With no DevTools MCP, measure layout shift in Playwright with a - Network accessline 92
- **Cost:** low to medium. Pick a tool the agent can run every time (Chrome DevTools MCP, Lighthouse CLI, axe, `curl -w`), and measure more than once when the number is noisy. With no DevTools MCP, measure layout shift in Playwright with a
skills/create-prove-it/references/skill-template.md
- Network accessline 75
| Flow or surface | Flow kind | Answer key | Rung | Where it lives | Written by | |---|---|---|---|---|---| | <e.g. MCP tools that copy app actions> | <new> | <the same action through the app> | <1> | <`curl localhost:$PORT/api/tasks`> | <t - Network accessline 75
| Flow or surface | Flow kind | Answer key | Rung | Where it lives | Written by | |---|---|---|---|---|---| | <e.g. MCP tools that copy app actions> | <new> | <the same action through the app> | <1> | <`curl localhost:$PORT/api/tasks`> | <t - Network accessline 76
|---|---|---|---|---|---| | <e.g. MCP tools that copy app actions> | <new> | <the same action through the app> | <1> | <`curl localhost:$PORT/api/tasks`> | <the existing app> | | <e.g. task list page> | <existing> | <screenshot> | <2> | <`. - Network accessline 76
|---|---|---|---|---|---| | <e.g. MCP tools that copy app actions> | <new> | <the same action through the app> | <1> | <`curl localhost:$PORT/api/tasks`> | <the existing app> | | <e.g. task list page> | <existing> | <screenshot> | <2> | <`. - Network accessline 77
| <e.g. MCP tools that copy app actions> | <new> | <the same action through the app> | <1> | <`curl localhost:$PORT/api/tasks`> | <the existing app> | - Network accessline 77
| <e.g. MCP tools that copy app actions> | <new> | <the same action through the app> | <1> | <`curl localhost:$PORT/api/tasks`> | <the existing app> |
How the signals fit together
Facts from a static read of the files at the pinned commit. They describe what the code does, not why.
Some files could not be read; the facts below still hold.
Signal confidence: 12 facts
- Network access in prose of documentationskills/create-prove-it/references/methods.md:90Low confidence
- Network access in prose of documentationcreate-prove-it/references/methods.md:90Low confidence
- Network access in prose of documentationskills/create-prove-it/references/methods.md:91Low confidence
- Network access in prose of documentationcreate-prove-it/references/methods.md:91Low confidence
- Network access in prose of documentationskills/create-prove-it/references/methods.md:92Low confidence
- Network access in prose of documentationcreate-prove-it/references/methods.md:92Low confidence
- Network access in code of documentationskills/create-prove-it/references/skill-template.md:75Low confidence
- Network access in code of documentationcreate-prove-it/references/skill-template.md:75Low confidence
- Network access in code of documentationskills/create-prove-it/references/skill-template.md:76Low confidence
- Network access in code of documentationcreate-prove-it/references/skill-template.md:76Low confidence
- Network access in code of documentationskills/create-prove-it/references/skill-template.md:77Low confidence
- Network access in code of documentationcreate-prove-it/references/skill-template.md:77Low confidence
Works with
| Agent | Status | Evidence | Install to |
|---|---|---|---|
| Claude Code | Native | Manifest .claude-plugin/marketplace.json | .claude/skills/<skill-name>/ |
| Codex | Compatible | Standard SKILL.md skills/audit-prove-it/SKILL.md | .agents/skills/<skill-name>/ |
| Cursor | Compatible | Standard SKILL.md skills/audit-prove-it/SKILL.md | .cursor/skills/<skill-name>/ |
| Gemini CLI | Compatible | Standard SKILL.md skills/audit-prove-it/SKILL.md | .gemini/skills/<skill-name>/ |
| GitHub Copilot | Compatible | Standard SKILL.md skills/audit-prove-it/SKILL.md | .github/skills/<skill-name>/ |
| OpenCode | Compatible | Standard SKILL.md skills/audit-prove-it/SKILL.md | .opencode/skills/<skill-name>/ |
- Native The repository ships a manifest, folder or frontmatter field for this agent.
- Compatible A standard SKILL.md that this agent can read.
- Unknown No evidence either way.
Derived automatically from paths, manifests and frontmatter. Not tested on each agent.
