Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add commands/bengous/claude-code-plugins/parallel-taskgit clone --depth 1 https://github.com/bengous/claude-code-pluginsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/bengous/claude-code-plugins/parallel-task)<a href="https://agentmods.dev/commands/bengous/claude-code-plugins/parallel-task"><img src="https://agentmods.dev/badge/commands/bengous/claude-code-plugins/parallel-task.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00012 | $0.00805 |
| Opus 5 | $0.00006 | $0.00402 |
| Sonnet 5 | $0.00002 | $0.00161 |
| Haiku 4.5 | $0.00001 | $0.00081 |
Grade A, and why
parallel-task scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 108 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Phase 0: Planning & Analysis
- Analyze the task and break into parallel work units
- Identify target scopes (folders/files to modify)
- Detect project verification commands:
- Check
package.jsonscripts (test, lint, typecheck, type-check) - Check for Makefile, pyproject.toml, Cargo.toml
- Note available commands for Phase 4
- Check
- Define safety requirements and verification criteria
- Present complete plan using AskUserQuestion for approval before execution
Phase 1: Exploration (Parallel)
Spawn N explore agents analyzing different areas.
Spawn ALL explore agents in a SINGLE message using parallel Agent tool calls:
Agent 1: subagent_type="Explore", model="sonnet", prompt="Analyze [scope1] for [task]. Report: files to change, patterns to follow, risks."
Agent 2: subagent_type="Explore", model="sonnet", prompt="Analyze [scope2] for [task]. Report: files to change, patterns to follow, risks."
...
sonnet is the floor for exploration work; do not route these to a cheaper tier.
Collect reports, identify what to change vs preserve, estimate impact.
Phase 2: Safety Setup
- Create rollback point:
git checkout -b backup/parallel-task-$(date +%s) - Add temporary files to .gitignore if needed
- Confirm all safety nets in place
Phase 3: Execution (Parallel)
Spawn N worker agents with strict scope isolation.
Spawn ALL worker agents in a SINGLE message using parallel Agent tool calls:
Agent 1: subagent_type="general-purpose", model="opus", prompt="[Detailed instructions for scope1]"
Agent 2: subagent_type="general-purpose", model="opus", prompt="[Detailed instructions for scope2]"
...
Each agent prompt includes:
- Explicit folder assignment (modify ONLY files in this scope)
- Step-by-step commands (not goals)
- Verification requirements (run lint/type-check before committing)
- Error handling: if blocked, return error report instead of partial work
- Commit changes with descriptive message including scope name
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday First seen · 108 lines · 12 tokens per session scan A 6b861ab2d535
parallel-task is a command published in the GitHub repository bengous/claude-code-plugins (4 stars, last pushed 3d ago), licensed MIT. It adds 12 tokens to every session and 805 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other commands, from other repositories
git-pr-push
Create git commit and push branch as PR using git + GitHub CLI.
pr-submit
Create git commit and submit current branch with Graphite (squashes commits and rebases stack).
commit
智能 Git 提交：分析改动生成 Conventional Commit 信息，支持拆分建议.
make-pr
Render a cognitive-aid PR body from flow-next state and open via gh.
commit
Create a git commit using the Haiku model.
start-work
Begin work on a Jira issue - creates branch, sets up environment.