token-config-checker
简体中文 | English
简体中文
用于检测包含 access_token / refresh_token / id_token 的 JSON 配置文件是否还有效。
提供内容
- -
scripts/check_token_configs.py:基础版,支持离线检测和可选在线探测 - INLINECODE4 :增强版,支持脱敏输出、失败原因分类、三分类保存、写入报告文件
快速用法
检测单个文件
CODEBLOCK0
批量检测目录
CODEBLOCK1
在线探测 access_token
CODEBLOCK2
增强版:按 valid / no_quota / invalid 三分类保存
CODEBLOCK3
输出重点
- -
valid_offline:离线结构和过期时间是否正常 - INLINECODE6 :在线探测是否通过
- INLINECODE7 :失败分类,例如
token_expired / INLINECODE9 - INLINECODE10 :最终分类(
valid / no_quota / invalid)
注意事项
- - 这类配置文件非常敏感,不要分享原始文件。
- 对 Codex 配置,在线结果通常比离线判断更重要。
- 默认探测地址使用 OpenAI 官方接口;如果你使用私有中转或自定义面板,请手动传入
--probe-url。 - 建议开启
--out 输出报告,便于复查。
English
Use this skill to validate JSON config files containing access_token / refresh_token / id_token.
Included scripts
- -
scripts/check_token_configs.py: basic version with offline checks and optional online probing - INLINECODE20 : enhanced version with redacted output, failure classification, three-bucket saving, and report generation
Quick usage
Check one file
CODEBLOCK4
Batch scan a directory
CODEBLOCK5
Online probe access_token
CODEBLOCK6
Enhanced: save configs into valid / no_quota / invalid buckets
CODEBLOCK7
Key outputs
- -
valid_offline: whether the file looks structurally valid offline - INLINECODE22 : whether online probing succeeded
- INLINECODE23 : failure reason classification
- INLINECODE24 : final bucket (
valid / no_quota / invalid)
Notes
- - These files are highly sensitive. Do not share raw configs.
- For Codex-style configs, online probe results are usually more important than offline checks.
- Default probe target uses the OpenAI official endpoint; if you use a private relay or custom panel, pass your own
--probe-url explicitly. - Prefer
--out to generate a reusable report.
token-config-checker
简体中文 | English
简体中文
用于检测包含 accesstoken / refreshtoken / id_token 的 JSON 配置文件是否还有效。
提供内容
- - scripts/checktokenconfigs.py:基础版,支持离线检测和可选在线探测
- scripts/checktokenconfigs_plus.py:增强版,支持脱敏输出、失败原因分类、三分类保存、写入报告文件
快速用法
检测单个文件
bash
python3 scripts/check
tokenconfigs.py /path/to/file.json
批量检测目录
bash
python3 scripts/check
tokenconfigs.py /path/to/dir
在线探测 access_token
bash
python3 scripts/check
tokenconfigs.py /path/to/file.json --probe
增强版:按 valid / no_quota / invalid 三分类保存
bash
python3 scripts/check
tokenconfigs_plus.py ./tokens \
--probe \
--probe-url https://api.openai.com/v1/models \
--save-valid-dir ./valid_configs \
--save-no-quota-dir ./no
quotaconfigs \
--save-invalid-dir ./invalid_configs \
--group-by-type \
--out report.txt
输出重点
- - validoffline:离线结构和过期时间是否正常
- onlinevalid:在线探测是否通过
- failureclass:失败分类,例如 tokenexpired / onlineinvalidorrevoked
- bucket:最终分类(valid / noquota / invalid)
注意事项
- - 这类配置文件非常敏感,不要分享原始文件。
- 对 Codex 配置,在线结果通常比离线判断更重要。
- 默认探测地址使用 OpenAI 官方接口;如果你使用私有中转或自定义面板,请手动传入 --probe-url。
- 建议开启 --out 输出报告,便于复查。
English
Use this skill to validate JSON config files containing accesstoken / refreshtoken / id_token.
Included scripts
- - scripts/checktokenconfigs.py: basic version with offline checks and optional online probing
- scripts/checktokenconfigs_plus.py: enhanced version with redacted output, failure classification, three-bucket saving, and report generation
Quick usage
Check one file
bash
python3 scripts/check
tokenconfigs.py /path/to/file.json
Batch scan a directory
bash
python3 scripts/check
tokenconfigs.py /path/to/dir
Online probe access_token
bash
python3 scripts/check
tokenconfigs.py /path/to/file.json --probe
Enhanced: save configs into valid / no_quota / invalid buckets
bash
python3 scripts/check
tokenconfigs_plus.py ./tokens \
--probe \
--probe-url https://api.openai.com/v1/models \
--save-valid-dir ./valid_configs \
--save-no-quota-dir ./no
quotaconfigs \
--save-invalid-dir ./invalid_configs \
--group-by-type \
--out report.txt
Key outputs
- - validoffline: whether the file looks structurally valid offline
- onlinevalid: whether online probing succeeded
- failureclass: failure reason classification
- bucket: final bucket (valid / noquota / invalid)
Notes
- - These files are highly sensitive. Do not share raw configs.
- For Codex-style configs, online probe results are usually more important than offline checks.
- Default probe target uses the OpenAI official endpoint; if you use a private relay or custom panel, pass your own --probe-url explicitly.
- Prefer --out to generate a reusable report.