How do I install the Beatlabs Harvester Harvester skill?

Run npx skillmds@latest add tomevault-io/beatlabs-harvester-harvester in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

What does the Beatlabs Harvester Harvester skill do?

fix-gh-issue Skill It is listed under Coding & Dev Tools on SkillMD.

Is Beatlabs Harvester Harvester safe to use?

This skill has not completed SkillMD's automated safety review yet. Independent scanners report: SkillSpector: PASS, Skill Scanner: PASS. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Beatlabs Harvester Harvester?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Beatlabs Harvester Harvester free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Beatlabs Harvester Harvester?

tomevault-io (@tomevault-io) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Beatlabs Harvester Harvester by tomevault-io

Name: Beatlabs Harvester Harvester
Author: tomevault-io

File contents

fix-gh-issue Skill

Orchestrate a single GitHub issue through the full lifecycle: worktree → implement → PR → CI → squash-merge → cleanup.

Works in any git repository with a GitHub remote.

Inputs

Extract from the user message:

ISSUE — GitHub issue number (required, e.g. 42)
BRANCH — branch name (optional — auto-derived if omitted)
PROMPT — custom prompt describing the fix (optional — default below)

If ISSUE is missing, ask before proceeding.

Derived values

Resolve at runtime — do not hardcode paths or repo names:

REPO_PATH=$(git rev-parse --show-toplevel)
REPO_NAME=$(basename "$REPO_PATH")

If BRANCH was not provided, derive it from the issue title:

BRANCH=$(gh issue view "$ISSUE" --json number,title \
  --jq '"fix/\(.number)-\(.title | gsub("^[a-z]+: "; "") | ascii_downcase | gsub("[^a-z0-9]+"; "-") | gsub("(^-|-$)"; "") | split("-")[0:6] | join("-"))"')

Then:

BRANCH_SAFE="${BRANCH//\//-}"
WORKTREE_PATH="${REPO_PATH}/../${REPO_NAME}.${BRANCH_SAFE}"

Default prompt (if none supplied):

Fix GitHub issue #<ISSUE> on branch <BRANCH>.
Run all tests before pushing.
Create a PR that closes #<ISSUE>.

Phase 1 — Create worktree

wt -C "$REPO_PATH" switch --create "$BRANCH" --no-cd

Skip if $WORKTREE_PATH already exists.

Phase 2 — Implement the fix

Work directly in $WORKTREE_PATH. Follow the prompt to implement, test, commit, and push. Create the PR:

gh pr create --title "..." --body "Closes #<ISSUE>" --head "$BRANCH"

Phase 3 — Wait for PR

If the PR was created by a separate agent or process, poll every 60 seconds (up to 2 hours):

gh pr list --head "$BRANCH" --json number --jq '.[0].number'

gh auto-detects the repo from the git remote. Log progress every 10 minutes. If no PR after 2 hours, stop and report.

Phase 4 — Wait for CI

gh pr checks "$PR_NUMBER" --watch --fail-fast

If CI fails: fix the issue in $WORKTREE_PATH, push, then re-run this phase.

Phase 5 — Squash-merge

gh pr merge "$PR_NUMBER" --squash --delete-branch

Phase 6 — Remove worktree

wt -C "$REPO_PATH" remove -D "$BRANCH"

Parallel execution

Each issue runs in its own isolated worktree — no filesystem conflicts. Run multiple issues in parallel freely. If two branches modify the same file, GitHub will flag the merge conflict on the PR when it's time to merge; deal with it then.

Integration test limitation: Integration tests share a single Consul and Redis instance with hardcoded keys. Running go test -tags=integration from two worktrees simultaneously will cause key collisions and flaky failures. Only one worktree should run integration tests at a time — coordinate manually or stagger the test runs.

Source: beatlabs/harvester — distributed by TomeVault.

---name: beatlabs-harvester-harvesterdescription: fix-gh-issue Skill---# fix-gh-issue SkillOrchestrate a single GitHub issue through the full lifecycle: worktree → implement → PR → CI → squash-merge → cleanup.Works in any git repository with a GitHub remote.## InputsExtract from the user message:- `ISSUE` — GitHub issue number (required, e.g. `42`)- `BRANCH` — branch name (optional — auto-derived if omitted)- `PROMPT` — custom prompt describing the fix (optional — default below)If `ISSUE` is missing, ask before proceeding.## Derived valuesResolve at runtime — do not hardcode paths or repo names:```bashREPO_PATH=$(git rev-parse --show-toplevel)REPO_NAME=$(basename "$REPO_PATH")```If `BRANCH` was not provided, derive it from the issue title:```bashBRANCH=$(gh issue view "$ISSUE" --json number,title \ --jq '"fix/\(.number)-\(.title | gsub("^[a-z]+: "; "") | ascii_downcase | gsub("[^a-z0-9]+"; "-") | gsub("(^-|-$)"; "") | split("-")[0:6] | join("-"))"')```Then:```bashBRANCH_SAFE="${BRANCH//\//-}"WORKTREE_PATH="${REPO_PATH}/../${REPO_NAME}.${BRANCH_SAFE}"```Default prompt (if none supplied):```Fix GitHub issue #<ISSUE> on branch <BRANCH>.Run all tests before pushing.Create a PR that closes #<ISSUE>.```## Phase 1 — Create worktree```bashwt -C "$REPO_PATH" switch --create "$BRANCH" --no-cd```Skip if `$WORKTREE_PATH` already exists.## Phase 2 — Implement the fixWork directly in `$WORKTREE_PATH`. Follow the prompt to implement, test, commit, and push. Create the PR:```bashgh pr create --title "..." --body "Closes #<ISSUE>" --head "$BRANCH"```## Phase 3 — Wait for PRIf the PR was created by a separate agent or process, poll every 60 seconds (up to 2 hours):```bashgh pr list --head "$BRANCH" --json number --jq '.[0].number'````gh` auto-detects the repo from the git remote. Log progress every 10 minutes. If no PR after 2 hours, stop and report.## Phase 4 — Wait for CI```bashgh pr checks "$PR_NUMBER" --watch --fail-fast```If CI fails: fix the issue in `$WORKTREE_PATH`, push, then re-run this phase.## Phase 5 — Squash-merge```bashgh pr merge "$PR_NUMBER" --squash --delete-branch```## Phase 6 — Remove worktree```bashwt -C "$REPO_PATH" remove -D "$BRANCH"```## Parallel executionEach issue runs in its own isolated worktree — no filesystem conflicts. Run multiple issues in parallel freely. If two branches modify the same file, GitHub will flag the merge conflict on the PR when it's time to merge; deal with it then.**Integration test limitation:** Integration tests share a single Consul and Redis instance with hardcoded keys. Running `go test -tags=integration` from two worktrees simultaneously will cause key collisions and flaky failures. Only one worktree should run integration tests at a time — coordinate manually or stagger the test runs.---> Source: [beatlabs/harvester](https://github.com/beatlabs/harvester) — distributed by [TomeVault](https://tomevault.io).

Frequently asked questions

Beatlabs Harvester Harvester