React Router Pull Request Creator
remix-run/react-router
Packages finished React Router work into a draft pull request: branch, commit, push, a written PR body and the right GitHub labels.
Commits a validated bug fix and publishes it in one of three modes: local-only, a workflow-owned handoff, or a direct pull request.
$ npx skills add omnigent-ai/omnigent --skill resolve-publish -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install omnigent-ai/omnigent resolve-publish --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/omnigent-ai/omnigent.git skills-src && mkdir -p .claude/skills && cp -r skills-src/dev/resolve-agent/skills/resolve-publish .claude/skills/resolve-publish && 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 "resolve-publish" agent skill from https://github.com/omnigent-ai/omnigent/tree/main/dev/resolve-agent/skills/resolve-publish into .claude/skills/resolve-publish/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "resolve-publish", 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/omnigent-ai/omnigent/tree/main/dev/resolve-agent/skills/resolve-publishType 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 omnigent-ai/omnigent --skill resolve-publish -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install omnigent-ai/omnigent resolve-publish --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/omnigent-ai/omnigent.git skills-src && mkdir -p .agents/skills && cp -r skills-src/dev/resolve-agent/skills/resolve-publish .agents/skills/resolve-publish && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "resolve-publish" agent skill from https://github.com/omnigent-ai/omnigent/tree/main/dev/resolve-agent/skills/resolve-publish into .agents/skills/resolve-publish/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "resolve-publish", 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 omnigent-ai/omnigent --skill resolve-publish -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install omnigent-ai/omnigent resolve-publish --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/omnigent-ai/omnigent.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/dev/resolve-agent/skills/resolve-publish .cursor/skills/resolve-publish && 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 "resolve-publish" agent skill from https://github.com/omnigent-ai/omnigent/tree/main/dev/resolve-agent/skills/resolve-publish into .cursor/skills/resolve-publish/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "resolve-publish", 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/omnigent-ai/omnigent.git --path dev/resolve-agent/skills/resolve-publish--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 omnigent-ai/omnigent --skill resolve-publish -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install omnigent-ai/omnigent resolve-publish --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/omnigent-ai/omnigent.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/dev/resolve-agent/skills/resolve-publish .gemini/skills/resolve-publish && 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 "resolve-publish" agent skill from https://github.com/omnigent-ai/omnigent/tree/main/dev/resolve-agent/skills/resolve-publish into .gemini/skills/resolve-publish/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "resolve-publish", 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 omnigent-ai/omnigent resolve-publishInstalls 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 omnigent-ai/omnigent --skill resolve-publish -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/omnigent-ai/omnigent.git skills-src && mkdir -p .github/skills && cp -r skills-src/dev/resolve-agent/skills/resolve-publish .github/skills/resolve-publish && 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 "resolve-publish" agent skill from https://github.com/omnigent-ai/omnigent/tree/main/dev/resolve-agent/skills/resolve-publish into .github/skills/resolve-publish/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "resolve-publish", 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 omnigent-ai/omnigent --skill resolve-publish -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install omnigent-ai/omnigent resolve-publish --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/omnigent-ai/omnigent.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/dev/resolve-agent/skills/resolve-publish .opencode/skills/resolve-publish && 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 "resolve-publish" agent skill from https://github.com/omnigent-ai/omnigent/tree/main/dev/resolve-agent/skills/resolve-publish into .opencode/skills/resolve-publish/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "resolve-publish", 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.
resolve-publishCommits a validated bug fix and publishes it in one of three modes: local-only, a workflow-owned handoff, or a direct pull request.
This is step 3 of a larger fix-resolution workflow, and it applies only when the agent wrote the fix itself. Before publishing, it repeats the earlier search for competing pull requests, inspects any new or changed candidates, rechecks whether earlier ones merged, and records the decision in fix_summary.
Three publication modes are defined. With skip_push set to true, it commits and stops. With an explicit CI publisher contract, it prepares and validates .omnigent/pr-body.md and the final handoff but makes no push or gh write. Otherwise it opens the PR directly. An unresolved design choice goes out as a draft through gh pr create --draft, with a partially_fixed handoff that lists the decision in remaining_work.
6 steps, taken from the first numbered list in SKILL.md.
Read from SKILL.md and the folder at commit 2e1cd15. 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:
ghgitpythonFrom 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:
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.
Bug Fix Publication Workflow loads about 4.5k tokens when it runs. Until then it costs about 26 tokens; SKILL.md has 2,515 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 omnigent-ai/omnigent at commit 2e1cd15, republished under its Apache-2.0 licence (© omnigent-ai). 2,515 words, ~4,526 tokens.
.claude/skills/resolve-publish/SKILL.md (or your agent's skills folder).This step applies only when you authored a fix in Step 2B — it's about opening a PR. (The review path 2A adopts the existing PR instead of opening one, then goes straight to Step 4 to land it.) Once the set is genuinely green:
Check again before publishing. Once the fix and PR body are ready, repeat
Step 1's search immediately before creating a new PR, or before the final
handoff to a CI publisher. Inspect only new or changed candidates, using Step 2A
if one may cover the bug; preserve your work while evaluating it. Recheck the
state of earlier candidates too: if one merged, use the shared repro audit on
updated main before deciding whether your fix is still needed. Record the check
and decision in fix_summary. Skip this refresh for skip_push and updates to
an existing PR.
skip_push: true) — commit the fix and stop at Step 3.2. No PR
will be published automatically, so do not prepare a PR body or run Step 4.
This takes precedence even when CI appended a generic publisher contract.skip_push: false plus an explicit CI publisher
contract) — do not push or make any gh write. Prepare and validate
.omnigent/pr-body.md using the body-writing instructions in Step 3.4, but do
not run its gh pr create command. Complete the deferred live-validation
preparation described in Step 4.4, then write the final handoff and stop. The
publisher performs the GitHub writes; do not run the PR-facing
CI/preview/review loop in the rest of Step 4.Unresolved design choice: after completing the investigation and available
validation in resolve-investigate, publish a direct author proposal with
gh pr create --draft. Explain the choice, evidence, alternatives, and remaining
checks in the body. Emit a partially_fixed handoff with the decision in
remaining_work. Skip preview, readiness checks, and the Step 4 review loop that
requires a ready PR; do not mark it ready merely to trigger automation. This is
a reviewable proposal, not proof the bug is fixed. For workflow-owned publication,
prepare the same body and incomplete handoff and obey the supplied publisher
contract; do not assume it supports drafts. skip_push still means local only.
In workflow-owned mode, .omnigent/ is intentionally gitignored, so body
transport does not rely on the file being committed. The workflow captures
pr-body.md separately in the resolve artifact bundle alongside the committed
checkpoint, then restores it into the publication worktree before running the PR
finalizer. The finalizer validates and uses that restored file as the PR
description; without it, the publisher can only construct a less readable
fallback from machine-oriented handoff fields.
When .omnigent/pr-gate.json exists, use the pr.py helper supplied in the
resolve-drive-pr skill directory. This mode replaces the direct GitHub write
and token-recovery recipes below. Keep Git pushes on the supplied credential;
never decode another token or change credential configuration.
| Action | Helper arguments |
|---|---|
| Open a PR after the required checks | create --title TITLE --body-file FILE --base main |
| Update its description | edit --number N --body-file FILE |
| Reply to findings | comment --number N --body-file FILE |
| Submit a review without approval | review --number N --head SHA --event COMMENT --body-file FILE |
| Mark a draft ready | ready --number N |
Run the helper from the fix checkout. Use review_cycle.py request for review
requests; CI routes it through the host. Follow CI's review budget and warning
policy. Actions the helper does not support remain maintainer actions; never
recover a broader credential, approve, or merge to work around that boundary.
The following token setup applies only when CI has not supplied the checked publication helper and has explicitly authorized direct publication.
Any write to GitHub — git push, gh pr create, gh pr edit --add-reviewer,
gh pr comment, gh pr close — needs the resolve-agent App installation token
(omni-resolve-agent[bot], Contents and Pull requests: write;
Actions: read and write on omnigent-ai/omnigent). Actions access is needed to
read review runs and completion artifacts and dispatch both review workflows.
Your shell does not inherit it in a usable env var:
you run inside the session's runner process (a different process, often a
different machine when hosted on --server), so $GH_TOKEN in your shell is
empty and a bare git push fails with a 403 / permission error. This is not
a missing/expired/read-only token — the write credential IS on this machine, in
the git config of your checkout. Recover it before any GitHub write.
The reliable source is the checkout's persisted http.extraheader.
actions/checkout bakes the App installation token into your repo's git config
as an AUTHORIZATION: basic <base64> header (git worktrees share it via the
common config, so it's readable from your fix worktree too). Decode it and export
it as GH_TOKEN:
# Run from anywhere inside your checkout / fix worktree. The extraheader value is
# base64("x-access-token:<token>"), so strip the prefix, base64 -d, take the part
# after the colon.
export GH_TOKEN="$(git config --get http.https://github.com/.extraheader \
| sed 's/^AUTHORIZATION: basic //' | base64 -d | cut -d: -f2-)"
[ -n "$GH_TOKEN" ] || echo "no extraheader token found in git config"
gh auth setup-git # route git pushes through gh's credential helper with this tokengh/git push call reports it lost auth). Then push, open the PR, request the
reviewer, and comment normally — all of them use this token. Confirm it works
and is write-scoped with gh auth status / a cheap gh api /repos/omnigent-ai/omnigent
before relying on it./proc/*/environ (that is denied in the session sandbox), and the ambient
github-actions[bot] credential is read-only on omnigent-ai/omnigent (it's
scoped to omnigent-internal) — both are dead ends that waste the turn. The
extraheader above is the one that works.skip_push run, or the
checkout didn't persist it), report that exact fact in maintainer_review with
the command output. Never substitute a guess like "token expired" or "PAT
is read-only" — those are false and drop the hand-off silently. Only a real,
quoted failure goes in maintainer_review.omnigent.forkPushTokenFile. That is a separate
maintainer credential for one purpose only: pushing a fix to an existing fork
PR whose author enabled maintainer edits. Never export it as GH_TOKEN and
never pass it to gh; PR creation, comments, reviews, labels, and every other
visible action must continue using the App token so GitHub attributes them to
omni-resolve-agent[bot].Once the set is genuinely green:
Commit the fix and selected permanent regression tests on the working
branch. Reused unchanged tests need no new commit. Follow the repo's commit
conventions. Before omitting an investigative test introduced for this task,
retain its original paths, command, tested revision, result, and source under
.omnigent/repro-evidence/, or cite the intact CI repro baseline/bundle.
Follow the evidence retention rules in 2B.4:
identify the CI run/artifact/path or verified persistent local copy, and
distinguish pending workflow upload from confirmed storage. A worktree-local
path alone is not durable; preserve that worktree if retention is unresolved.
Put selected committed tests in tests and the selection/evidence location in
test_audit; no reproduction-only source is required in the PR. This does not
authorize deleting existing repository tests solely to reduce LOC.
Never commit workspace artifacts. In particular, never stage or commit
the recordings/ clips or any .omnigent/ handoff files (e.g.
.omnigent/repro-handoff.json): recordings are workspace artifacts that ride
in the PR's Demo section / CI artifact bundle, not in the diff (see
dev/recording-lanes.md). Do not use a blanket
git add -A / git add . that sweeps them in — stage the fix and test paths
explicitly, and run git status / git diff --cached --stat before committing
to confirm the staged set is only the fix + test. If a recording or handoff
file already landed in an earlier commit on this branch, remove it (e.g.
git rm --cached) so it never reaches the PR.
Read the staged diff for redundant comments and docstrings, including tests
carried over from repro. Keep only short explanations of non-obvious
constraints; move investigation history to the handoff. If a safety
explanation runs long, consider clearer code or a focused docstring first.
Refresh the shared impact assessment against the committed deliverable. If a
hook changed tested files, rerun their affected checks before the handoff.
After committing, if the target checkout has the advisory checker, run
python .github/scripts/pr-template/hygiene.py --base origin/main (using the
target's default branch). Review any long added comment blocks and trim only
redundant text. If the checker is unavailable, inspect the diff manually;
the workflow-owned publisher runs its own copy before creating a PR.
If the input has skip_push: true, stop here — the fix is committed
locally; do not push and do not open a PR. Report the branch name in
your output (pushed_branch) so a human can inspect, push, and PR it. The
focused local validation in 2B.5 still runs before the handoff is written.
This local-only input also suppresses workflow-owned publication; never treat
the presence of the generic CI publisher overlay as permission to continue.
Otherwise push the branch. First make sure git push / gh have the
write token — see "Get the GitHub write token" below. Your shell does not
inherit GH_TOKEN (you run in the session's runner, not the CI wrapper's
process), so echo $GH_TOKEN is normally empty and a bare git push / gh pr create fails with a permission error. Recover the token first; do not
conclude the token is "expired" or "read-only" from an empty env var — it is
present on the machine, just not exported to your shell.
Open a ready-for-review PR with gh pr create, unless the draft proposal
exception above applies. Automated review runs on ready PRs. Create
.omnigent/ if needed. If the target repository provides
.github/pull_request_template.md, copy it to
.omnigent/pr-body.md and edit that file. Otherwise create
.omnigent/pr-body.md with concise Related issue, Summary, and Test
Plan sections. Pass the finished file to gh pr create --body-file .omnigent/pr-body.md. The
workflow-owned publisher also restores this file from the resolve artifact
bundle if it has to finish publication after your session ends, so write it
before the GitHub call or final handoff. Link the bug in the template's
Related issue section.
Write the description for a reviewer, not for the handoff parser:
Root Cause, Validation, or Issues
sections.resolve-investigate; a passing test alone does
not justify the policy.facets, test_transition, other handoff fields, or a long comma-
separated inventory of test names into the body.If the target repository provides the template validator, validate the body locally before publishing it:
PR_BODY="$(cat .omnigent/pr-body.md)" \
python .github/scripts/pr-template/validate.pyFix every validation error before gh pr create. In Related issue, use a
GitHub closing keyword only against a GitHub issue number —
Resolve #<closing_issue_number> (equivalently Closes #<n>), using the
closing_issue_number you determined in Step 1 (the bug_url issue, or the
mirrored GitHub issue for a Linear ticket). Never point a closing keyword
at a raw Linear URL — GitHub can't close it, and it clutters the body. When
there is no closing_issue_number (Linear-only bug with no mirror), don't use
a closing keyword at all: reference the ticket in prose
(e.g. "Resolves OMNI-1234 (Linear)"). Then summarize the root cause and the
fix, and in the Test Plan give the concrete fail→pass proof (test paths,
the pre-fix fail reason, the post-fix pass). Check
"Bug fix" and the test-coverage boxes that apply. Generate the body from the
actual diff and this reproduction — do not skip template sections. Put the
before/after recordings in the Demo section: upload the files when your
environment can attach media to the PR; otherwise link where they live (the
CI run's artifact bundle, or the repro session) so reviewers can watch the
failure and the fix. For internal/API-only results with no visible user
interaction, put the written before/after evidence in Demo. If recording
was blocked, explain why and include the available evidence. When the bug
is a Linear ticket and a Linear key is available, also attach both recordings
to the ticket (GraphQL fileUpload + attachmentCreate).
If the target checkout has .github/scripts/pr-template/hygiene.py, run it
on the committed diff and finished body before publication:
python .github/scripts/pr-template/hygiene.py \
--base origin/main --body-file .omnigent/pr-body.mdUse the target's default branch when it is not main. Review warnings about
long comments, body length, or repeated prose; trim redundant text while
keeping necessary safety explanations. Warnings are advisory and do not
replace template validation. If the checker is unavailable, review the body
manually; the workflow-owned publisher runs its own copy before PR creation.
Emit an interim handoff now — the moment the PR is open. As soon as
gh pr create succeeds, print the full handoff json block (the Output schema)
with pr_url set and outcome at its current best assessment, before you
start Step 4. This is what lets the workflow post the PR link to the Linear
ticket promptly, rather than waiting the ~hour Step 4 can take. Leave the
not-yet-known Step-4 fields empty (ci_status, polly_review, ocr_review,
maintainer_review, with review_cycle: {}) — refill them in the final
handoff. Emit it as a normal intermediate message (json block last in that message), then carry on.
For a draft proposal, emit the incomplete handoff and finish without the
outward actions below or Step 4. Otherwise, before this handoff, do the two
outward actions a mid-turn drop would otherwise strand:
ui-preview (author path) — gh pr edit <pr> --add-label ui-preview. Your own PR is same-repo, already-pipelined code, so it needs no
CI-green gate (see 4.1); label it now so the preview builds while you drive
Step 4. (Review-path fork PRs still wait for green — 4.1.)Supersedes #<old> on its own
line, and set reviewed_pr_url so workflow reconciliation can preserve and
inform the original PR until the replacement merges.You do not merge. For a ready PR, go to Step 4 and drive it to a green, reviewed, ready-for-a-human state. Draft proposals use the exception above.
© omnigent-ai, Apache-2.0. 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 dev/resolve-agent/skills/resolve-publish of omnigent-ai/omnigent.
Open the folder on GitHubat commit 2e1cd15
Bug Fix Publication Workflow 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 |
|---|---|---|---|---|---|---|
| Bug Fix Publication Workflow this skillomnigent-ai/omnigent | 11k | — | ~4.5k | Automated safety check: Pass | Apache-2.0 | |
| React Router Pull Request Creatorremix-run/react-router | 57k | — | ~2.5k | Automated safety check: Pass | MIT | |
| Draft Pull Request Creatorwordpress-mobile/WordPress-Android | 3.2k | — | ~881 | Automated safety check: Notes | GPL-2.0 | |
| Codewhale Landing Workflowcodewhale-hq/Codewhale | 41k | — | ~1.6k | Automated safety check: Pass | MIT | |
| PR Preparerflutter-ml/google_ml_kit_flutter | 1.3k | — | ~1.2k | Automated safety check: Pass | MIT | |
| Publish Pull Request for taskctltaskctl/taskctl | 434 | — | ~979 | Automated safety check: Pass | GPL-3.0 |
remix-run/react-router
Packages finished React Router work into a draft pull request: branch, commit, push, a written PR body and the right GitHub labels.
wordpress-mobile/WordPress-Android
Commits and pushes current changes, writes a pull request title and body from the branch history and template, and opens a draft PR on GitHub after you approve it.
codewhale-hq/Codewhale
Decides how verified work should reach main, directly, in a worktree or on an integration branch, while keeping contributor credit and respecting merge gates.
flutter-ml/google_ml_kit_flutter
Analyzes git branch diffs (added, modified, deleted files) against the default or target branch, and generates a conventional commit title, PR title, and formatted markdown PR description for…
taskctl/taskctl
Publishes the current branch and opens or updates a GitHub pull request for the taskctl repo, after a pre-publish gate and with a description template matched to the change.
c5inco/compose-pokedexer
Handles Pokedexer Git and GitHub workflows: inspect changes, prepare commit messages, manage branches and pushes, and create or update issues and pull requests with safe file-based inputs.
omnigent-ai/omnigent
Brings up the Omnigent server and Postgres as a Docker compose stack on any Docker host, and covers the Dockerfile's runtime and host build targets for extending it to a new platform.
omnigent-ai/omnigent
Scans Python agent code for framework imports and recommends the matching Omnigent executor type, or says when the framework is not natively supported yet.
omnigent-ai/omnigent
Runs the Omnigent load test with real hosts and multi-turn sessions against a mocked LLM, then explains the latency results from summary.md.
omnigent-ai/omnigent
Spins up an isolated Omnigent server, runner and mock model to prove a user-facing behavior or bug fix with recorded evidence instead of reasoning from code.
omnigent-ai/omnigent
Spins up a local Omnigent server and exercises the Antigravity (Gemini) SDK harness end to end: building agents, running real turns, smoke tests and bug-bashing.
omnigent-ai/omnigent
Gives patterns for generating a minimal, valid Omnigent agent directory: the config.yaml fields, the right executor type, and the files each agent needs.
Works with
Categories
Commits a validated bug fix and publishes it in one of three modes: local-only, a workflow-owned handoff, or a direct pull request. This is step 3 of a larger fix-resolution workflow, and it applies only when the agent wrote the fix itself. Before publishing, it repeats the earlier search for competing pull requests, inspects any new or changed candidates, rechecks whether earlier ones merged, and records the decision in fix_summary.
Bug Fix Publication Workflow fits situations like: committing a verified fix after tests and checks pass; deciding whether to push a PR or leave publication to a CI publisher; publishing a draft PR when a design choice is still open; rechecking for duplicate pull requests before opening one.
Run `npx skills add omnigent-ai/omnigent --skill resolve-publish -a claude-code`. Or copy the skill folder (dev/resolve-agent/skills/resolve-publish in omnigent-ai/omnigent) into .claude/skills/resolve-publish in your project. Claude Code loads it when a task matches its description.
Run `npx skills add omnigent-ai/omnigent --skill resolve-publish -a codex`. Or copy the skill folder (dev/resolve-agent/skills/resolve-publish in omnigent-ai/omnigent) into .agents/skills/resolve-publish 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 omnigent-ai/omnigent --skill resolve-publish -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/resolve-publish, .gemini/skills/resolve-publish, .github/skills/resolve-publish and .opencode/skills/resolve-publish in your project.
Going by SKILL.md and its folder, Bug Fix Publication Workflow needs the command-line tools its instructions call (gh, git and python) and credentials named GH_TOKEN. Our summary lists: Git and the GitHub CLI (gh); A validated fix from the earlier investigation step.
SKILL.md names 1 domain. In commands or code: 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.
Bug Fix Publication Workflow is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
About 4.5k tokens (SKILL.md is roughly 18k 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 Bug Fix Publication Workflow: React Router Pull Request Creator (remix-run/react-router, 57k stars), Draft Pull Request Creator (wordpress-mobile/WordPress-Android, 3.2k stars), Codewhale Landing Workflow (codewhale-hq/Codewhale, 41k stars) and PR Preparer (flutter-ml/google_ml_kit_flutter, 1.3k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
omnigent-ai (a GitHub organization) maintains it in omnigent-ai/omnigent, which has 10,691 GitHub stars. The repository holds 19 skills in this directory. The repository was last updated on October 9, 2026.
Source: omnigent-ai/omnigent on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.