[aw] Prompt Optimization reported incomplete result

Đang mở
  • Có gì không ổn?
#333 0 bình luận 0 reaction 0 người được giao Xem trên GitHub

Maintainer thường phản hồi trong vòng 1 ngày

Chưa có ai nhận issue này.

Đánh giá

Độ khó
4/5
Thời gian dự kiến
3-5 ngày
Mức phù hợp với người mới
42/100
Loại issue
Lỗi
Độ rõ ràng
Khá rõ ràng
Mức độ hoạt động
Sôi nổi
Công nghệ
github-actions, ollama
Lĩnh vực
devops, infrastructure

Hướng nghiên cứu

Start with .github/workflows/prompt-optimization.lock.yml around the Ollama setup and model pull steps cited in the issue, then read .github/workflows/prompt-optimization.md and the linked workflow run. Trace why Ollama and its models are unavailable inside the agent sandbox and check how the workflow exposes the service and network access. Done when the workflow can run its optimization evaluation and report a result.

Do mô hình lập chỉ mục viết ra từ nội dung của issue.

Mô tả

agentic-workflows
Workflow Failure

Workflow: Prompt Optimization
Branch: main
Run: https://github.com/githubnext/gh-aw-wizard/actions/runs/38310132548

[!WARNING]
Task Could Not Be Completed: The agent reported that the task could not be performed due to an infrastructure or tool failure.

Reasons:

  • The Ollama proxy-model infrastructure this task depends on is not reachable from the agent sandbox, so no hill-climbing evaluation could be run.

    Pre-flight check: searched for an open PR titled "Prompt optimization:" via gh pr list — none exists, so optimization should have proceeded.

Infrastructure failure: (127.0.0.1/redacted) (the Ollama server the task says is already running) was not listening in this sandbox — curl returned connection refused (exit 7) from the start, before I touched anything. No ollamaprocess was running and~/.ollama` did not exist (no cached blobs/models restored into this container).

Recovery attempt 1: Located the ollama binary at /opt/hostedtoolcache/ollama/0.33.2/x64/ollama and started ollama serve myself in-container. It came up fine (Listening on 127.0.0.1:11434, total blobs: 0), confirming no model cache was present here.

Recovery attempt 2: Tried ollama pull hf.co/bartowski/Qwen2.5-1.5B-Instruct-GGUF:Q4_K_M to populate it locally. The outbound firewall (squid proxy) returned 403 Forbidden for both hf.co and huggingface.co — these hosts are not in this workflow's allowed egress domains, so the model cannot be fetched from inside the agent sandbox. Also probed likely host-gateway addresses (docker bridge gateway, 172.30.0.1, host.docker.internal) in case the actual GitHub Actions job (which per the lock.yml does actions/cache restore + ai-action/setup-ollama + ollama pull with full default/node network access) had a reachable instance; none responded, and allow-host-ports: [11434] did not expose any live host-side listener to this container.

I stopped my own locally-started ollama serve process and reverted an unrelated auto-generated metadata diff in .github/skills/optimize-scenario-prompt/SKILL.md that appeared during skill restoration, so the working tree is clean (git status --porcelain empty). No candidate was measured, no files were changed, and no pull request or review comment was created.

Likely cause: the GitHub Actions job steps that set up Ollama and pulled the two GGUF models (per .github/workflows/prompt-optimization.lock.yml, lines ~503-528) run in the outer GitHub-hosted runner, but this agent executes inside the stricter docker-sudo-iptables sandbox, whose egress allowlist (defaults, node, local) does not include hf.co/huggingface.co, and the host's already-running Ollama server (if started) was not port-forwarded into this container despite allow-host-ports: [11434].

This is a structured incompletion signal (report_incomplete), not a real task outcome. Any other safe outputs emitted alongside this signal (e.g., comments) describe the failure state, not a completed review or action.

Action Required

Assign this issue to an agent to debug and fix the issue.

Debug with any coding agent

Use this prompt with any coding agent (GitHub Copilot, Claude, Gemini, etc.):

Debug the agentic workflow failure using https://raw.githubusercontent.com/github/gh-aw/main/debug.md

The failed workflow run is at https://github.com/githubnext/gh-aw-wizard/actions/runs/38310132548
Manually invoke the agent

Debug this workflow failure using your favorite Agent CLI and the agentic-workflows prompt.

[!TIP]

Stop reporting this workflow as a failure

To stop a workflow from creating failure issues, set report-failure-as-issue: false in its frontmatter:

safe-outputs:
  report-failure-as-issue: false

Generated from Prompt Optimization · copilot · 100.3 AIC · ◷

Ngôn ngữ chính
JavaScript
Star
6
Fork
1
Merge trung bình
16 giờ 15 phút
Pull request đã merge (30 ngày)
16

Chuẩn bị môi trường

Dự án này không cung cấp dev container, Dockerfile hay hướng dẫn đóng góp, nên bạn cần tự thiết lập môi trường: hãy bắt đầu từ README và xem hướng dẫn đóng góp lần đầu của chúng tôi để biết các bước chung.

Bắt đầu từ đâu

  1. Đọc hết issue, rồi đọc hướng dẫn đóng góp của dự án.
  2. Bình luận trên issue rằng bạn sẽ nhận — tránh hai người làm cùng một việc.
  3. Fork repository và làm thay đổi trên một nhánh.
  4. Mở pull request có tham chiếu số hiệu của issue.

Issue khác của githubnext/gh-aw-wizard

Tất cả issue của githubnext/gh-aw-wizard

Issue tương tự

Thêm issue về JavaScript

Nhận issue mới trong hộp thư của bạn

Bản tóm tắt ngắn những issue GitHub phù hợp với người mới.