| name | krx-auth-recovery |
| description | Use when official KRX-backed pykrx paths fail in eit-market-data. Trigger on KRX login changes,
`KrxAuthRequired`, `LOGOUT`, `400`, `KeyError('지수명')`, empty KRX index or listing frames,
pykrx dataframe column errors like missing `종가`/`BPS`/`PER`,
browser cookie/profile issues, or when scripts/krx_login.py and src/eit_market_data/kr/krx_auth.py
need to be modified or debugged.
|
KRX Auth Recovery
Use this skill for official KRX session problems, not for general CI-safe bundle work.
Workflow
- Confirm whether the failing path is truly KRX-authenticated or a CI-safe path.
- Inspect
krx_auth.py, market_helpers.py, pykrx_provider.py, and scripts/krx_login.py.
- Check whether the installed
pykrx has pykrx.website.comm.auth.KRXSession.
- Reproduce with
scripts/preflight_kr_data.py before changing code.
- Preserve the rule that GitHub-hosted CI must not depend on KRX browser login.
- If the user only needs KR research data in CI, redirect to
kr-bundle-pipeline.
Modern pykrx Rule
Recent pykrx versions can authenticate directly from KRX_ID and KRX_PW.
When that auth module exists, do not install legacy webio.Get/Post monkey-patches.
Overriding webio can bypass pykrx's authenticated session and make KRX POSTs return
LOGOUT even when credentials are valid.
Use ensure_krx_authenticated_session(interactive=False) as the local authenticated
probe. A successful non-interactive path should make scripts/krx_login.py --timeout 30
print [OK] probe rows=... without needing the browser.
All official pykrx scripts should load project .env before pykrx calls. If this is missing,
crawl_kr_data_pykrx.py can fail with misleading column errors (종가, BPS, PER,
지수명) even though the real issue is missing KRX_ID/KRX_PW.
Read Next
- Known symptoms and recovery steps:
references/runbook.md
- CI-safe alternative path:
../kr-bundle-pipeline/SKILL.md
Commit Format
If you make a commit while using this skill, follow @rules/commit-messages.md:
<type>: <message> then a blank line then <body>.
