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/bifrost-proxy/bifrost/github-actions-patnpx skills add bifrost-proxy/bifrost --skill github-actions-patgit clone --depth 1 https://github.com/bifrost-proxy/bifrostWhat 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 | $0.00109 | $0.03217 |
| Opus 5 | $0.00055 | $0.01608 |
| Sonnet 5 | $0.00022 | $0.00643 |
| Haiku 4.5 | $0.00011 | $0.00322 |
Grade A, and why
github-actions-pat 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 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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
如果用户本机挂了代理(包括 bifrost 自己),Python 的 `urllib` 可能因 `Missing Authority Key Identifier` 而 `CERTIFICATE_VERIFY_FAILED`。跑脚本时 **主动清掉代理环境变量**: How it starts
The opening of the file, as written. The whole thing — 202 lines — stays where its author put it; the contents beside it link to each section on GitHub.
GitHub Actions PAT Inspector
把"用 PAT 分析 GitHub Actions CI"这条路径固化成可复用脚本。本 skill 解决的典型场景:
- Agent 在 bifrost remote 目标机或 caller 上跑,必须走 REST API 自动化,不走浏览器 cookie
- 需要自动化「修 → push → 盯 CI → 失败继续修」的闭环
- 需要按 run-id / PR / sha / branch 精确定位失败并拉 job log
- 需要做 PR diff review(生成 markdown 或直接 POST)
auth 唯一路径:PAT 从 GITHUB_TOKEN 环境变量读取。不支持、不回退到 cookie / OAuth device flow / SSH / gh CLI 登录等其他方式。需要 agent 自主跑的 CI 分析、PR review、fix-push-watch 循环,全部用本 skill。
授权契约(唯一输入:GITHUB_TOKEN 环境变量)
- 本 skill 只从环境变量
GITHUB_TOKEN读 token。不落盘、不写日志、不回显原文。 - 读取方式推荐(适配用户 shell 配置):
# macOS / zsh 用户:交互式 shell 才会加载 ~/.zshrc zsh -ic 'python3 scripts/gh_ci.py run <run_id>' # bash 用户 bash -lc 'python3 scripts/gh_ci.py run <run_id>' # 或在脚本运行前已 export GITHUB_TOKEN 的环境 python3 scripts/gh_ci.py run <run_id> - 未设置 → 脚本
exit 2,提示ERROR: set GITHUB_TOKEN before running this skill。 - 推荐 scope:
repo+actions:read(只读分析)。--post发 review 时额外需要pull_requests:write。 - 首选 fine-grained PAT 并把 repo 限定到目标仓库,爆炸半径最小。
仓库定位策略
脚本默认读环境变量 GH_REPO(格式 owner/repo),未设置时回退到本仓库。Agent 可在调用前写入:
export GH_REPO=bifrost-proxy/bifrost
python3 scripts/gh_ci.py pr 567
或者在 AGENTS.md 顶部的仓库元信息中固化。
目录结构
scripts/
├── common.py # token/http/分页/日志切片/归因
├── gh_ci.py # run / pr / sha / branch / regression
├── gh_review.py # PR metadata + diff + 分层建议 + 可选 --post
└── poll_run.py # 轮询一个 run 直到完成
references/
└── pitfalls.md # GitHub API 坑点清单(job log 302、Accept 415、system proxy MITM)
典型调用
(A) CI 错误定位
# 按 run-id 分析
python3 scripts/gh_ci.py run 25269751068
# 按 PR 号找最近一次 failed
python3 scripts/gh_ci.py pr 567
# 按 commit sha 找
python3 scripts/gh_ci.py sha a96a4257
# 按分支找最近一次 failed
python3 scripts/gh_ci.py branch feat/agent --only-failed
# 与上一次 green 做 regression 对比(给出 compare URL + 提交区间)
python3 scripts/gh_ci.py regression 25269751068
脚本行为:
GET /repos/{owner}/{repo}/actions/runs/...拉 run 元数据GET .../jobs筛出conclusion=failure- 对每个 failed job:
GET .../jobs/{id}/logs→ 自动处理 302 到 Azure Blob signed URL(不能带 Authorization) - 日志末尾 500 行扫关键词(
error:、FAIL、panicked、##[error]、thread '...' panicked、test result: FAILED等),抽 ±20 行上下文 - 输出结构化 markdown:失败 job / step / URL / 根因桶(compile / test / lint / fmt / timeout / network / OOM)/ 关键日志片段 / 本地复现命令
What ships with it
7 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.
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 · 202 lines · 109 tokens per session scan A e950fbe8e7b8
github-actions-pat is a skill published in the GitHub repository bifrost-proxy/bifrost (124 stars, last pushed 3d ago), licensed MIT. It adds 109 tokens to every session and 3,217 once invoked, about $0.0005 per session on Opus 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 2026-08-30.
Other skills, from other repositories
aws-solution-architect
Design AWS architectures for startups using serverless patterns and IaC templates. Use when asked to design serverless architecture, create CloudFormation templates, optimize AWS costs, set up CI/CD pipelines, or migrate to AWS. Covers Lambda, API Gateway, DynamoDB, ECS, Aurora, and cost optimization.
cc-devops-skills
SRE, DevOps, Kubernetes, CI/CD, PromQL, Terraform, Docker, and incident operations playbook for building reliable delivery and operations workflows.
ci-cd-and-automation
Automates CI/CD pipeline setup. Use when setting up or modifying build and deployment pipelines. Use when you need to automate quality gates, configure test runners in CI, or establish deployment strategies.
github
Interact with GitHub using the gh CLI to manage repositories, issues, pull requests, CI/CD workflow runs, and API queries. Use when the user asks to create, list, view, merge, or close pull requests and issues; check CI status or workflow run logs; query the GitHub API for repository data; or perform any GitHub…
baby-sit
Monitor a GitHub pull request until CI is green, diagnose failures, and rerun only evidence-backed flaky GitHub Actions jobs.
pipeline-investigation
Investigates Buildkite pipeline failures to find root causes. Returns structured JSON to the parent for formatting. Triggers when users ask about failing pipelines, build errors, or need help debugging CI/CD issues. Accepts Buildkite build URLs or build numbers and performs deep investigation.