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 skills/valpere/chorus/parallel-debugnpx skills add valpere/chorus --skill parallel-debuggit clone --depth 1 https://github.com/valpere/chorusWhat 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 | $0.00044 | $0.00478 |
| Opus 5 | $0.00022 | $0.00239 |
| Sonnet 5 | $0.00009 | $0.00096 |
| Haiku 4.5 | $0.00004 | $0.00048 |
Grade A, and why
chorus-parallel-debug 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 2d ago.
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.
What it actually says
Chorus: Parallel Debug
When to use
- Root cause of a bug is unclear
- User says "parallel debug", "chorus debug", "hypotheses from all agents"
Your role
You generate hypotheses focused on edge cases in input handling, type coercion, off-by-one errors.
Invocation
SYMPTOM="<symptom>"
claude --print "Root-cause hypotheses, ranked by likelihood (numbered list). Focus: application logic, state management.\n\nSymptom: $SYMPTOM" --dangerously-skip-permissions &
CLAUDE_PID=$!
gemini --prompt "Root-cause hypotheses, ranked by likelihood (numbered list). Focus: infrastructure, concurrency, environment.\n\nSymptom: $SYMPTOM" --yolo --output-format text &
GEMINI_PID=$!
agent -p --force "Root-cause hypotheses, ranked by likelihood (numbered list). Focus: framework, library, and third-party integration issues.\n\nSymptom: $SYMPTOM" &
CURSOR_PID=$!
kilo run --auto "Root-cause hypotheses, ranked by likelihood (numbered list). Focus: naming confusion, type misuse, readability issues that hide bugs.\n\nSymptom: $SYMPTOM" &
KILO_PID=$!
wait $CLAUDE_PID $GEMINI_PID $CURSOR_PID $KILO_PID
Graceful degradation: Check each agent with command -v <binary> before spawning. Skip missing agents and warn the user. Proceed as long as at least 2 agents (including yourself) are available.
Output handling
Merge all available hypothesis lists (up to four agents + your edge-case hypotheses), prioritize ones flagged by multiple agents:
## Hypothesis Pool (ranked)
1. … (Claude + Gemini)
2. … (Codex)
## Investigation Plan
1. Check X first — rules out hypotheses 1 and 3
Known limitation
Codex sandbox limits access to the current working directory.
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.
- 2d ago First seen · 55 lines · 44 tokens per session scan A c567293bb8ad
chorus-parallel-debug is a skill published in the GitHub repository valpere/chorus (10 stars, last pushed 3mo ago), licensed MIT. It adds 44 tokens to every session and 478 once invoked, about $0.0002 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 2026-08-31.
Other skills, from other repositories
engram-rebuild
이미 설치된 engram 환경에 개발 변경을 적용한다. dev-rebuild.ps1 로 충분한지 INSTALL.ps1 전체를 돌려야 하는지 판단하고 실행한다. 트리거 — 재빌드, 리빌드, dev 변경 적용, 설치본에 반영, overlay 재시작, exe 갱신, "고친 거 확인하려면 뭐 돌려야 해", rebuild, redeploy.
orchestrate
Run Engram's Orchestrator-Planner-Coder-Servant workflow for multi-step development, refactors, or research-plus-implementation tasks that need an acceptance contract and independent verification. Trigger for "orchestrate", delegation requests, planning then implementation, or complex multi-step development. Do not…
engram-new-session
Skill "engram-new-session" from JJHbrams/Project-AMBER, covering engram 말풍선 — 새 세션 시작, 실행 절차 and 규칙.
dead-code-cleanup
Identifies and removes dead/orphaned code, tests, and Storybook stories — including trivial tests that check only obvious behavior or rare edge cases with no real value. Always validates that code is truly unused before deleting anything. Trigger on: "remove dead code", "clean up unused files", "delete orphaned…
issue-maintenance
Periodic and on-demand maintenance pass over a repo's open GitHub issue backlog. Re-labels miscategorized or unlabeled issues (bug/enhancement/question/invalid/duplicate) and bug lifecycle labels (needs-repro/needs-info) following the same decision logic as Anthropic's own dogfooded triage-issue command; re-scores…
visual-crawl
Crawls the running app at randomized breakpoints, screenshots visual issues, checks design token consistency, and tests interactive elements. Each run covers different ground by design. Trigger on: "visual QA", "audit the app visually", "screenshot all the pages", "check for visual regressions", "crawl the site for…