PR Babysitter
openinterpreter/openinterpreter
Watches an open GitHub pull request until it merges, handling review comments, diagnosing CI failures and retrying flaky checks along the way.
Process GitHub issues as an automation queue: /gh-issues lists and filters issues, confirms selected issue numbers, deduplicates fix/issue- work, delegates one focused PR per issue, and can monitor…
$ npx skills add HybridAIOne/hybridclaw --skill gh-issues -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install HybridAIOne/hybridclaw gh-issues --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/HybridAIOne/hybridclaw.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/gh-issues .claude/skills/gh-issues && 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 "gh-issues" agent skill from https://github.com/HybridAIOne/hybridclaw/tree/main/skills/gh-issues into .claude/skills/gh-issues/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "gh-issues", 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/HybridAIOne/hybridclaw/tree/main/skills/gh-issuesType 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 HybridAIOne/hybridclaw --skill gh-issues -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install HybridAIOne/hybridclaw gh-issues --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/HybridAIOne/hybridclaw.git skills-src && mkdir -p .agents/skills && cp -r skills-src/skills/gh-issues .agents/skills/gh-issues && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "gh-issues" agent skill from https://github.com/HybridAIOne/hybridclaw/tree/main/skills/gh-issues into .agents/skills/gh-issues/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "gh-issues", 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 HybridAIOne/hybridclaw --skill gh-issues -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install HybridAIOne/hybridclaw gh-issues --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/HybridAIOne/hybridclaw.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/skills/gh-issues .cursor/skills/gh-issues && 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 "gh-issues" agent skill from https://github.com/HybridAIOne/hybridclaw/tree/main/skills/gh-issues into .cursor/skills/gh-issues/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "gh-issues", 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/HybridAIOne/hybridclaw.git --path skills/gh-issues--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 HybridAIOne/hybridclaw --skill gh-issues -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install HybridAIOne/hybridclaw gh-issues --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/HybridAIOne/hybridclaw.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/skills/gh-issues .gemini/skills/gh-issues && 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 "gh-issues" agent skill from https://github.com/HybridAIOne/hybridclaw/tree/main/skills/gh-issues into .gemini/skills/gh-issues/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "gh-issues", 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 HybridAIOne/hybridclaw gh-issuesInstalls 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 HybridAIOne/hybridclaw --skill gh-issues -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/HybridAIOne/hybridclaw.git skills-src && mkdir -p .github/skills && cp -r skills-src/skills/gh-issues .github/skills/gh-issues && 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 "gh-issues" agent skill from https://github.com/HybridAIOne/hybridclaw/tree/main/skills/gh-issues into .github/skills/gh-issues/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "gh-issues", 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 HybridAIOne/hybridclaw --skill gh-issues -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install HybridAIOne/hybridclaw gh-issues --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/HybridAIOne/hybridclaw.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/skills/gh-issues .opencode/skills/gh-issues && 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 "gh-issues" agent skill from https://github.com/HybridAIOne/hybridclaw/tree/main/skills/gh-issues into .opencode/skills/gh-issues/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "gh-issues", 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.
gh-issuesProcess GitHub issues as an automation queue: /gh-issues lists and filters issues, confirms selected issue numbers, deduplicates fix/issue- work, delegates one focused PR per issue, and can monitor…
Gh Issues is an agent skill from HybridAIOne/hybridclaw. Process GitHub issues as an automation queue: /gh-issues lists and filters issues, confirms selected issue numbers, deduplicates fix/issue- work, delegates one focused PR per issue, and can monitor issue-fix PR review feedback. Use this instead of general PR workflow tools only when the entry point is a GitHub issue list or issue-fix queue.
Its SKILL.md is about 4.7k 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 Development, covering Pull requests. It works with GitHub. The repository describes itself as: Enterprise-ready self-hosted AI assistant runtime with sandboxed execution, secure credentials, approvals, and memory. The licence is MIT.
6 steps, taken from the step headings in SKILL.md.
Read from SKILL.md and the folder at commit 3624072. It shows what the files ask for, not the result of running them.
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.
Shell commands in SKILL.md call:
ghgitFrom the folder's file list and the shell code blocks in SKILL.md.
Hosts in commands or code, which the agent is likely to contact:
api.github.comFrom URLs in SKILL.md, links to its own repository left out.
Names these keys or tokens, usually read from environment variables:
GH_TOKENFrom names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Gh Issues loads about 4.7k tokens when it runs. Until then it costs about 88 tokens; SKILL.md has 1,923 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.
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.
The full file from HybridAIOne/hybridclaw at commit 3624072, republished under its MIT licence (© HybridAIOne). 1,923 words, ~4,676 tokens.
.claude/skills/gh-issues/SKILL.md (or your agent's skills folder).You are an issue queue orchestrator. Follow the phases in order. Do not run processing preflight before the user has selected issues.
Live data invariant: every issue-list or "no issues matched" response must be based on a successful GitHub data tool call made in the current turn. Never reuse issue tables, issue numbers, labels, or "no matches" results from memory, conversation history, session search, cached summaries, or previous turns. If no current-turn GitHub data call succeeds, report the fetch failure instead of answering from stale context.
Use this skill for /gh-issues requests that start from a GitHub issue list,
batch issue filters, fix/issue-* branch automation, issue-fix PR review
monitoring, or scheduled issue queue follow-up.
Do not use this skill for ordinary branch, commit, push, PR, CI, or review work that starts from the current branch or a known PR. Use a GitHub PR workflow or code-review skill for those.
Parse the arguments after /gh-issues.
Positional:
owner/repo: source repository. If omitted, infer it from
git remote get-url origin; if that is unavailable, ask for owner/repo.
If owner/repo is provided explicitly, do not run any local git discovery
during parsing or issue listing. Local checkout checks belong only to Phase 4
after issue selection.Flags:
| Flag | Default | Description |
|---|---|---|
--label <label> | none | Filter by label. |
--limit <n> | 10 | Max issues to fetch per poll. |
--milestone <milestone> | none | Filter by milestone title. |
--assignee <assignee> | none | Filter by assignee; resolve @me with gh api user --jq .login. |
--state <open|closed|all> | open | Issue state. |
--fork <owner/repo> | none | Push branches to a fork while PRs target the source repo. |
--watch | false | Fetch the issue list normally, then schedule recurring issue and review follow-up after the first confirmed run. |
--interval <minutes> | 5 | Watch interval; only valid with --watch. |
--cron | false | Recurring-run mode: process at most one eligible item and exit. |
--dry-run | false | Fetch and display issues only. |
--yes | false | Process listed issues without another confirmation. |
--reviews-only | false | Skip issue fetching and process issue-fix PR review feedback. |
--model <model> | none | Optional model override for delegated tasks. |
--notify-channel <target> | none | Optional HybridClaw message target for final summaries only. |
Derived values:
SOURCE_REPO: issue and PR target repository.PUSH_REPO: --fork value, otherwise SOURCE_REPO.PUSH_OWNER: owner portion of PUSH_REPO.PUSH_REMOTE: fork when --fork is used, otherwise origin.FORK_MODE: true when --fork is used.SOURCE_REPO_SLUG: SOURCE_REPO with / replaced by -.Mode routing:
--reviews-only is set, run authentication, then jump to Phase 6.--cron is set, force --yes.--watch is set, use --interval, defaulting to 5 minutes. Do not skip
issue fetching or confirmation. The first watch turn follows Phases 2 and 3
like a normal run.GitHub CLI is preferred but optional. First check whether it exists:
command -v ghIf gh exists, prefer GitHub CLI authentication:
gh auth statusIf gh is unavailable or authentication fails, use the GitHub REST API through
the http_request tool with bearerSecretName: "GH_TOKEN". The gateway resolves
GH_TOKEN from the encrypted secret store; do not ask the shell for GH_TOKEN,
do not echo tokens, and do not use curl for authenticated GitHub API calls. If
neither gh auth nor stored secret GH_TOKEN works, ask the user to run
gh auth login or store a GitHub token in this order: browser admin at
/admin/secrets, /secret set GH_TOKEN <token> in
browser /chat or TUI, then local console fallback
hybridclaw secret set GH_TOKEN <token>.
When not in --reviews-only, fetch issues with a current-turn GitHub data call.
Primary gh path:
gh issue list --repo "$SOURCE_REPO" --state "$STATE" --limit "$LIMIT" \
--json number,title,labels,assignees,url,bodyAdd optional filters only when present:
--label "$LABEL"--milestone "$MILESTONE"--assignee "$ASSIGNEE" after resolving @meAPI fallback path with gh still available:
gh api "repos/$SOURCE_REPO/issues" \
-f state="$STATE" -f per_page="$LIMIT"API fallback path without gh: call http_request with
bearerSecretName: "GH_TOKEN" and the GitHub URL. Example request shape:
{
"method": "GET",
"url": "https://api.github.com/repos/{SOURCE_REPO}/issues?state={STATE}&per_page={LIMIT}",
"bearerSecretName": "GH_TOKEN",
"headers": {
"Accept": "application/vnd.github+json"
}
}Add API query parameters only when present:
labels=$LABELassignee=$ASSIGNEE after resolving @me through GET /usermilestone=$MILESTONE_NUMBER; if the user supplied a title, first list
milestones and match the title to a numbergh issue list excludes pull requests. API issue endpoints include pull
requests, so exclude any item with a pull_request field.
Hard failure rule: if the current turn did not execute a successful gh issue list, GitHub Issues API, or equivalent http_request call, do not display an
issue table and do not say no issues matched. Report that live issue fetch did
not complete.
If watch context includes PROCESSED_ISSUES, filter those issue numbers out.
If no issues match, report that directly and, in watch mode, continue to Phase 6
to check review feedback.
Extract for each issue: number, title, body, label names, assignees, URL.
Display a compact table:
| # | Title | Labels | Assignees |
|---|
If fork mode is active, also say that branches will be pushed to PUSH_REPO
and PRs will target SOURCE_REPO.
If --dry-run is set, stop after the table. Do not run preflight.
If --yes is set, process every listed issue and proceed to Phase 4.
Otherwise ask:
Which issues should I process:
all, a comma-separated list of issue numbers, orcancel?
Wait for the user response. Continue only with selected issue numbers.
Watch behavior: ask on the first interactive poll unless --yes is set. Do not
schedule watch follow-up from a parse-only turn. After the user confirms issues
or after a --yes run completes, schedule the next run in Watch Mode.
Subsequent scheduled runs use --cron --yes.
Run this phase only after selected issue numbers are known. These checks are for delegating fixes, not for issue listing.
Resolve the local checkout.
Use the current directory if it is a git repo whose origin matches
SOURCE_REPO. If the current directory is not a matching checkout, ask the
user for the local checkout path before processing. Issue listing does not
require a local checkout, but fixing issues does.
Check the working tree.
git status --porcelainIf output is non-empty, explain that delegated fixes should start from a committed state and ask whether to continue. Never stage or commit unrelated changes.
Record the base branch.
git rev-parse --abbrev-ref HEADStore as BASE_BRANCH.
Verify remotes.
git ls-remote --exit-code origin HEADIn fork mode, verify a fork remote exists for PUSH_REPO; add it only if
missing, and use the normal GitHub remote URL. Do not write tokens into git
config.
Skip issues that already have open PRs from the intended branch owner.
gh pr list --repo "$SOURCE_REPO" --state open \
--head "fix/issue-$ISSUE_NUMBER" \
--json number,url,headRefName,headRepositoryOwner \
--jq ".[] | select(.headRepositoryOwner.login == \"$PUSH_OWNER\")"If gh is unavailable, use http_request with
bearerSecretName: "GH_TOKEN":
{
"method": "GET",
"url": "https://api.github.com/repos/{SOURCE_REPO}/pulls?head={PUSH_OWNER}:fix/issue-{ISSUE_NUMBER}&state=open&per_page=1",
"bearerSecretName": "GH_TOKEN",
"headers": {
"Accept": "application/vnd.github+json"
}
}Skip issues whose intended branch exists in PUSH_REPO.
BRANCH_REF="fix%2Fissue-$ISSUE_NUMBER"
gh api "repos/$PUSH_REPO/branches/$BRANCH_REF" --silentIf gh is unavailable, use http_request with
bearerSecretName: "GH_TOKEN" against:
https://api.github.com/repos/{PUSH_REPO}/branches/fix%2Fissue-N.
Track claims for cron/watch dedupe when durable state storage is available.
Use $HYBRIDCLAW_STATE_DIR/gh-issues when that variable is set, otherwise
$HOME/.hybridclaw/data/gh-issues. Create the directory if needed. Store
claims in claims.json, creating it as {} if missing. Remove claims older
than 2 hours. For each remaining issue, skip if key
SOURCE_REPO#ISSUE_NUMBER is claimed. Never write claim or cursor files into
the target repository checkout.
If all selected issues are skipped, report why and, in watch mode, continue to Phase 6.
Before each delegate call, write or refresh the claim for that issue when the claims file is available. Include enough context for the delegated task to work without reading this orchestration transcript.
Cron mode:
cursor-SOURCE_REPO_SLUG.json in the same HybridClaw state directory as
the claims file.{"last_processed":null,"in_progress":null} if missing.last_processed when possible. Wrap to the first eligible issue if needed.in_progress, delegate one issue, report the spawn, and exit. Do not
wait for results.Normal mode:
delegate in parallel mode for up to 6 independent issues at a time.Delegate task shape:
You are a focused issue-fix subagent. Fix exactly one GitHub issue and open a PR.
Repository: {SOURCE_REPO}
Push repository: {PUSH_REPO}
Fork mode: {FORK_MODE}
Push remote: {PUSH_REMOTE}
Base branch: {BASE_BRANCH}
Branch: fix/issue-{NUMBER}
Notification target: {NOTIFY_CHANNEL}
Issue #{NUMBER}: {TITLE}
URL: {URL}
Labels: {LABELS}
Assignees: {ASSIGNEES}
Body:
{BODY}
Workflow:
1. Confirm the issue is actionable. If confidence is below 7/10, stop and
report why instead of guessing.
2. Create branch `fix/issue-{NUMBER}` from `{BASE_BRANCH}`.
3. Find the smallest relevant code or docs change.
4. Implement only the issue fix.
5. Run targeted tests or explain why no targeted test exists.
6. Stage only your changes and commit with:
`fix: {SHORT_DESCRIPTION}`
Include `Fixes {SOURCE_REPO}#{NUMBER}` in the commit body.
7. Push to `{PUSH_REMOTE}` and open a pull request against `{SOURCE_REPO}`.
8. If a notification target is present, send only the final PR summary to that
target with the message tool; do not send status chatter.
9. Return: PR URL, files changed, tests run, skipped checks, and caveats.
Constraints:
- Do not modify unrelated files.
- Do not force-push.
- Do not commit secrets, local config, or personal data.
- If the working tree contains unrelated changes, leave them untouched.
- If the issue is vague, too large, or not reproducible, return analysis and
mark it for manual triage instead of opening a speculative PR.Result collection:
| Issue | Status | PR | Notes |
|---|
Statuses: PR opened, Skipped, Failed, Timed out, Needs manual review.
Store opened PRs as OPEN_PRS with PR number, branch, URL, and issue number for
Phase 6.
If --notify-channel is set, send a final summary only after the batch
finishes. Do not include issue bodies, review bodies, tokens, or transcripts.
This phase monitors open issue-fix PRs for actionable review feedback.
It runs:
--reviews-only is setDiscover PRs:
OPEN_PRS exists from Phase 5, inspect those first.fix/issue-:gh pr list --repo "$SOURCE_REPO" --state open \
--json number,title,url,headRefName,body \
--jq '.[] | select(.headRefName | startswith("fix/issue-"))'Fetch review sources for each candidate:
gh pr view "$PR_NUMBER" --repo "$SOURCE_REPO" \
--json number,title,url,body,reviews,comments,reviewDecision
gh api "repos/$SOURCE_REPO/pulls/$PR_NUMBER/comments"
gh api "repos/$SOURCE_REPO/issues/$PR_NUMBER/comments"Also inspect the PR body for embedded review content such as
<!-- greptile_comment --> or structured sections from review bots.
Determine the bot username with:
gh api user --jq .loginExclude comments authored by that user and comments already answered with addressed-style replies.
Actionable feedback includes:
CHANGES_REQUESTED reviewsNot actionable:
Build ACTIONABLE_COMMENTS with source, author, body, path, line, diff hunk,
comment ID, and URL when available.
If no actionable comments are found, report that and stop unless watch mode needs scheduling.
Display a table:
| PR | Branch | Actionable Comments | Sources |
|---|
Unless --yes, --cron, or a subsequent scheduled watch run is active, ask
which PRs to address: all, comma-separated PR numbers, or skip.
Cron review mode:
Normal review mode:
delegate in parallel mode for up to 6 PRs at a time.Review delegate task shape:
You are a focused PR review-fix subagent. Address actionable review feedback on
one pull request.
Repository: {SOURCE_REPO}
Push repository: {PUSH_REPO}
Push remote: {PUSH_REMOTE}
PR #{PR_NUMBER}: {PR_URL}
Branch: {BRANCH}
Actionable comments:
{COMMENTS}
Workflow:
1. Fetch and checkout `{BRANCH}` from `{PUSH_REMOTE}`.
2. Read every actionable comment and group changes by file.
3. Implement only requested review fixes.
4. Run targeted tests or explain why none apply.
5. Commit with `fix: address review feedback on PR #{PR_NUMBER}`.
6. Push the branch.
7. Reply to addressed comments with a brief summary and commit SHA. For
unresolved or contradictory comments, reply with the reason and mark manual
follow-up.
8. Return: comments addressed, comments skipped, commit SHA, files changed,
tests run, and manual follow-ups.
Constraints:
- Do not force-push.
- Do not rewrite unrelated PR history.
- Do not change files unrelated to the actionable comments.
- If comments conflict, address the newest concrete request and flag the
conflict.After review delegates finish, present:
| PR | Comments Addressed | Comments Skipped | Commit | Status |
|---|
Track addressed comment IDs in ADDRESSED_COMMENTS for watch context.
HybridClaw must not sleep in the active turn. If --watch is set, use cron
to schedule a recurring follow-up prompt using the same repository and filters.
The initial --watch invocation is not a parse-only request. It must fetch the
current issue list first, present the table, and either ask for selected issue
numbers or honor --yes. Only schedule the recurring follow-up after that first
selection/processing path has run.
Use the cron tool for scheduled follow-up. Do not send a local message as a
substitute for scheduling. If cron is unavailable or scheduling fails, report
that watch scheduling failed.
The scheduled prompt must include:
SOURCE_REPO, PUSH_REPO, FORK_MODE, PUSH_REMOTE--cron --yes--reviews-only when the run should only monitor review feedbackPROCESSED_ISSUES, ADDRESSED_COMMENTS, OPEN_PRS,
cumulative one-line results, and BASE_BRANCH when knownThe scheduled prompt must not include:
If a watch run finds no issue work, still run Phase 6. If no issue or review work exists, report "No eligible issue or review work found" and let the next scheduled run handle future polling.
© HybridAIOne, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Just SKILL.md in skills/gh-issues of HybridAIOne/hybridclaw.
Open the folder on GitHubat commit 3624072
Gh Issues 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 |
|---|---|---|---|---|---|---|
| Gh Issues this skillHybridAIOne/hybridclaw | 158 | — | ~4.7k | Automated safety check: Pass | MIT | |
| PR Babysitteropeninterpreter/openinterpreter | 69k | 3 repos | ~4.2k | Automated safety check: Pass | Apache-2.0 | |
| Check PRonyx-dot-app/onyx | 32k | 2 repos | ~2.3k | Automated safety check: Pass | MIT | |
| Contributor-First PR MergeHKUDS/OpenHarness | 16k | 1 repos | ~847 | Automated safety check: Pass | MIT | |
| Create Pull Requestcline/cline | 70k | 1 repos | ~1.6k | Automated safety check: Pass | Apache-2.0 | |
| Pull Request Title and Body Writeropeninterpreter/openinterpreter | 69k | 2 repos | ~1.1k | Automated safety check: Pass | Apache-2.0 |
openinterpreter/openinterpreter
Watches an open GitHub pull request until it merges, handling review comments, diagnosing CI failures and retrying flaky checks along the way.
onyx-dot-app/onyx
Checks a GitHub, GitLab, or Perforce (p4) pull request (or merge request, or shelved changelist) for unresolved review comments, failing status checks, and incomplete PR descriptions.
HKUDS/OpenHarness
Merges external GitHub pull requests while keeping the original author credited, and fixes conflicts after the merge instead of rewriting the contribution.
cline/cline
Opens a GitHub pull request from your current branch with the gh CLI, after reviewing the commits and diff and gathering the details the PR needs.
openinterpreter/openinterpreter
Rewrites the title and body of one or more pull requests with gh, leading with why the change was made, then what changed, and describing only the net result.
prisma/orm
Runs a loop on a GitHub pull request: fetch review state, triage comments into actions, implement them and resolve threads, repeating until nothing actionable is left.
HybridAIOne/hybridclaw
Use Hermes3000 to plan, draft, revise, save, check consistency, and export long-form manuscripts through the Hermes3000 AI writing portal API.
HybridAIOne/hybridclaw
Plan, script, render, and stitch Manim Community Edition videos in Python.
HybridAIOne/hybridclaw
Create and update SKILL.md-based skills with strong trigger metadata, lean docs, and reliable init, validate, package, and publish workflows.
HybridAIOne/hybridclaw
Create, edit, inspect, and analyze .xlsx spreadsheets and Excel workbooks.
HybridAIOne/hybridclaw
Create and revise editable .excalidraw diagrams as Excalidraw JSON for architecture diagrams, flowcharts, sequence diagrams, concept maps, and other hand-drawn explainers.
HybridAIOne/hybridclaw
Manage Google Ads accounts with safe GAQL reporting, campaign planning, guarded mutations, and gateway-proxied REST API calls.
Works with
Categories
Process GitHub issues as an automation queue: /gh-issues lists and filters issues, confirms selected issue numbers, deduplicates fix/issue- work, delegates one focused PR per issue, and can monitor…. Gh Issues is an agent skill from HybridAIOne/hybridclaw. Process GitHub issues as an automation queue: /gh-issues lists and filters issues, confirms selected issue numbers, deduplicates fix/issue- work, delegates one focused PR per issue, and can monitor issue-fix PR review feedback.
Gh Issues fits situations like: tasks that involve Pull requests.
Run `npx skills add HybridAIOne/hybridclaw --skill gh-issues -a claude-code`. Or copy the skill folder (skills/gh-issues in HybridAIOne/hybridclaw) into .claude/skills/gh-issues in your project. Claude Code loads it when a task matches its description.
Run `npx skills add HybridAIOne/hybridclaw --skill gh-issues -a codex`. Or copy the skill folder (skills/gh-issues in HybridAIOne/hybridclaw) into .agents/skills/gh-issues in your project. Codex loads it when a task matches its description.
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add HybridAIOne/hybridclaw --skill gh-issues -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/gh-issues, .gemini/skills/gh-issues, .github/skills/gh-issues and .opencode/skills/gh-issues in your project.
Going by SKILL.md and its folder, Gh Issues needs the command-line tools its instructions call (gh and git) and credentials named GH_TOKEN.
SKILL.md names 1 domain. In commands or code: api.github.com; the agent is likely to contact it when it follows the instructions. This is read from the text; nothing was executed.
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.
Gh Issues is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
About 4.7k tokens (SKILL.md is roughly 19k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.
Skills that share tags, products or a category with Gh Issues: PR Babysitter (openinterpreter/openinterpreter, 69k stars), Check PR (onyx-dot-app/onyx, 32k stars), Contributor-First PR Merge (HKUDS/OpenHarness, 16k stars) and Create Pull Request (cline/cline, 70k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
HybridAIOne (a GitHub organization) maintains it in HybridAIOne/hybridclaw, which has 158 GitHub stars. The repository holds 71 skills in this directory. The repository was last updated on October 9, 2026.
Source: HybridAIOne/hybridclaw on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.