Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/inclusionAI/Avernetnpx agentmods add skills/inclusionai/avernet/clawevolve-diagnoseWrote 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/skills/inclusionai/avernet/clawevolve-diagnose)<a href="https://agentmods.dev/skills/inclusionai/avernet/clawevolve-diagnose"><img src="https://agentmods.dev/badge/skills/inclusionai/avernet/clawevolve-diagnose/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/inclusionai/avernet/clawevolve-diagnose"><img src="https://agentmods.dev/badge/skills/inclusionai/avernet/clawevolve-diagnose.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00117 | $0.03914 |
| Opus 5.5 | $0.00047 | $0.01566 |
| Sonnet 5.5 | $0.00023 | $0.00783 |
| Haiku 4.5 | $0.00012 | $0.00391 |
Grade A, and why
clawevolve-diagnose scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 21d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
3. **只运行一次脚本**：禁止先空跑 `bash scripts/run.sh`，禁止失败后换参数重跑，禁止手工 curl ClawWeb 代替脚本。 How it starts
The opening of the file, as written. The whole thing — 216 lines — stays where its author put it; the contents beside it link to each section on GitHub.
clawevolve-diagnose
平台核心实现选择
按下文运行原诊断脚本。脚本在会话分析调用处选择默认或自定义业务实现，继续负责采集、筛选、产物生成与上报；Agent 无需另行调度或提交结果。
最高优先级执行规则
- 用户怎么调用，就怎么传给脚本：不要改参数名、不要改 ID 字符、不要补参数、不要删除参数。
- 只运行一次脚本：禁止先空跑
bash scripts/run.sh，禁止失败后换参数重跑，禁止手工 curl ClawWeb 代替脚本。 - ID 必须原样保留：
EV-...、STEP-...中的连字符-不能改成下划线_；参数名必须是--task-id、--step-id。
执行前自检最终 shell 命令：
- shell 命令必须包含用户原始
--task-id、--step-id；若用户提供了--api-key、--model等参数，也必须逐字符保留。 - shell 命令中的 task-id/step-id 必须与用户文本逐字符一致。
角色定位
收到 /clawevolve-diagnose ... 或用户要求“诊断历史 session / 抽取 eval case / 挖掘 good/bad case / 为 clawevolve-plan 准备输入”时，Agent 只做脚本执行器：运行脚本、等待结束、读取最终 stdout JSON、按 JSON 指示下一步。不要手工分析 session、不要重写流程、不要自行上传或生成模板。
固定链路：发现本地 sessions → selected session judge 诊断 → 去重选 case → 写 artifacts/plan-source.json → 上报 ClawWeb step report。模板生成属于 clawevolve-plan，不在本 skill 内完成。
必须执行的命令
收到 /clawevolve-diagnose ... 后，只运行一次前台脚本，并把用户的完整 slash command 作为一个单独引号参数传给 scripts/run.sh。这是唯一推荐方式，可最大程度避免自然语言、日期范围和 ID 被拆坏：
cd clawevolve-diagnose && bash scripts/run.sh '<用户完整/clawevolve-diagnose命令原文>'
稳定脚本调用建议使用显式 --intent：
cd clawevolve-diagnose && bash scripts/run.sh \
--task-id "$TASK_ID" \
--step-id "$STEP_ID" \
--judge-backend subagent \
--intent "诊断工具调用失败和异步任务未完成的问题，抽取最多10个case，重点关注最近7天"
示例：用户输入是：
/clawevolve-diagnose --api-key *** --model GLM-5 --intent "抽取 20260723 至 20260804 的5个case，有good和bad，bad占比至少80%。" --task-id EV-20260804-444491B4 --step-id STEP-A70F0B24A0804F33
则执行：
cd clawevolve-diagnose && bash scripts/run.sh '/clawevolve-diagnose --api-key *** --model GLM-5 --intent "抽取 20260723 至 20260804 的5个case，有good和bad，bad占比至少80%。" --task-id EV-20260804-444491B4 --step-id STEP-A70F0B24A0804F33'
注意：不要把示例中的 *** 当真实 key。--judge-backend subagent 使用当前 Bot 的 OpenClaw Agent，
不需要 API Key；--judge-backend api 才要求 --api-key 或 OPENAI_API_KEY。未显式选择 backend 时，
有 Key 兼容走 API，否则默认走 subagent。ARCA Message 模式只允许 subagent，不得在 Message 中传 Key。
What ships with it
60 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- agents/openai.yaml 306 B
- clawevolve_diagnose/__init__.py 41 B runs code
- clawevolve_diagnose/_ocsa_session_report/__init__.py 35 B runs code
- clawevolve_diagnose/_ocsa_session_report/domain/__init__.py 35 B runs code
- clawevolve_diagnose/_ocsa_session_report/domain/features.py 13 KB runs code
- clawevolve_diagnose/_ocsa_session_report/evaluators/__init__.py 35 B runs code
- clawevolve_diagnose/_ocsa_session_report/evaluators/session_report/__init__.py 35 B runs code
- clawevolve_diagnose/_ocsa_session_report/evaluators/session_report/evaluator.py 33 KB runs code
- clawevolve_diagnose/_ocsa_session_report/evaluators/session_report/prompt.py 22 KB runs code
- clawevolve_diagnose/_ocsa_session_report/evaluators/session_report/schema.py 15 KB runs code
- clawevolve_diagnose/_ocsa_session_report/infrastructure/__init__.py 35 B runs code
- clawevolve_diagnose/_ocsa_session_report/infrastructure/llm/__init__.py 13 KB runs code
- clawevolve_diagnose/_ocsa_session_report/infrastructure/odps/__init__.py 35 B runs code
- clawevolve_diagnose/_ocsa_session_report/infrastructure/odps/adapter.py 15 KB runs code
- clawevolve_diagnose/_ocsa_session_report/infrastructure/odps/writer.py 20 KB runs code
- clawevolve_diagnose/_ocsa_session_report/parsers/__init__.py 35 B runs code
- clawevolve_diagnose/_ocsa_session_report/parsers/conversation_parser.py 12 KB runs code
- clawevolve_diagnose/_ocsa_session_report/parsers/normalizers.py 18 KB runs code
- clawevolve_diagnose/_ocsa_session_report/prompts/__init__.py 35 B runs code
- clawevolve_diagnose/_ocsa_session_report/prompts/common.py 4.3 KB runs code
- clawevolve_diagnose/_ocsa_session_report/utils/__init__.py 35 B runs code
- clawevolve_diagnose/_ocsa_session_report/utils/type_utils.py 5.8 KB runs code
- clawevolve_diagnose/acquisition/__init__.py 58 B runs code
- clawevolve_diagnose/acquisition/discovery.py 3.7 KB runs code
- clawevolve_diagnose/acquisition/runtime_identity.py 7.1 KB runs code
- clawevolve_diagnose/acquisition/service_export.py 41 KB runs code
- clawevolve_diagnose/acquisition/sessions.py 58 KB runs code
- clawevolve_diagnose/acquisition/tracing.py 3.6 KB runs code
- clawevolve_diagnose/artifacts/__init__.py 58 B runs code
- clawevolve_diagnose/artifacts/case_ids.py 540 B runs code
- clawevolve_diagnose/artifacts/outputs.py 1.7 KB runs code
- clawevolve_diagnose/artifacts/reporting.py 11 KB runs code
- clawevolve_diagnose/cli.py 7.1 KB runs code
- clawevolve_diagnose/constants.py 6.9 KB runs code
- clawevolve_diagnose/core.py 3.5 KB runs code
- clawevolve_diagnose/integration/__init__.py 58 B runs code
- clawevolve_diagnose/integration/clawweb_events.py 12 KB runs code
- clawevolve_diagnose/integration/output.py 10 KB runs code
- clawevolve_diagnose/integration/plan_source.py 9.9 KB runs code
- clawevolve_diagnose/intent.py 27 KB runs code
- clawevolve_diagnose/judge/__init__.py 58 B runs code
- clawevolve_diagnose/judge/agent_session_contract.py 2.2 KB runs code
- clawevolve_diagnose/judge/analyzer_factory.py 1.6 KB runs code
- clawevolve_diagnose/judge/keyless_session_analyzer.py 11 KB runs code
- clawevolve_diagnose/judge/keyless_session_prompt.py 1.4 KB runs code
- clawevolve_diagnose/judge/local_session_judge_provider.py 39 KB runs code
- clawevolve_diagnose/judge/native_session_analysis.py 15 KB runs code
- clawevolve_diagnose/judge/ocsa_contract.py 299 B runs code
- clawevolve_diagnose/judge/ocsa_labels.py 3.2 KB runs code
- clawevolve_diagnose/judge/ocsa_request_matcher.py 25 KB runs code
- clawevolve_diagnose/judge/ocsa_session_adapter.py 19 KB runs code
- clawevolve_diagnose/judge/ocsa_session_report_analyzer.py 40 KB runs code
- clawevolve_diagnose/judge/openai_chat_client.py 30 KB runs code
- clawevolve_diagnose/judge/openclaw_subagent_client.py 51 KB runs code
- clawevolve_diagnose/judge/query_fidelity.py 4.1 KB runs code
- clawevolve_diagnose/judge/query_rewriter.py 25 KB runs code
- clawevolve_diagnose/judge/runtime.py 3.2 KB runs code
- clawevolve_diagnose/judge/workspace_defaults/AGENTS.md 605 B
- clawevolve_diagnose/judge/workspace_defaults/BOOTSTRAP.md 358 B
- clawevolve_diagnose/judge/workspace_defaults/HEARTBEAT.md 66 B
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.
- 21d ago Changed · +6 lines 730f75b37136
- 23d ago Changed · +1 lines b107aad9c5c1
- 1mo ago Changed dc6ad984c265
- 1mo ago First seen · 209 lines · 117 tokens per session scan A 2b8431f3bd55
clawevolve-diagnose is a skill published in the GitHub repository inclusionAI/Avernet (641 stars, last pushed today), licensed Apache-2.0. It adds 117 tokens to every session and 3,914 once invoked, about $0.0005 per session on Opus 5.5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-15.
Other skills, from other repositories
investigate-sentry-issue
Investigate and triage a Sentry error issue.
sre-agent
Deploy or update the SRE Agent on Omnara, a production investigator that diagnoses issues across AWS and source code and reports the likely cause with evidence. Use when the user asks to deploy, set up, update, or remove the SRE agent.
apps
Build and iterate on schema-backed Swarm Apps with live models, named queries, custom actions, and json-render interfaces using the app-upsert, app-get, app-list, and app-patch tools.
kapso-whatsapp
How to interact with Kapso WhatsApp from the swarm — read inbound webhook payloads (text AND media), fetch message history, send free-form messages within the 24h session window (and template messages outside it), mark-as-read, show the typing indicator, send reactions, download media, verify webhook signatures, and…
swarm-scripts
Bulk, repeat, fan-out, or data-heavy work: write and run swarm scripts (inline script-run, named script-upsert, durable launch-script-run). Covers the script-vs-tool rubric, the authoring contract (args first, ctx second), the seed catalog, connections and secrets, dbquery, and exposing a script as an API.
composio
Use Composio from Agent Swarm through the agent-swarm x composio CLI route, the swarmx MCP tool, or a registered ctx.api.composio script connection. Trigger when a task needs connected third-party app tools such as Gmail, Google Calendar, Google Docs, Google Drive, GitHub, Slack, Notion, or HubSpot through Tool Router…