[Anuj7411]

v1.7.5: read compressed Codex logs

Codex now compresses its older session logs, and Sipcode reads them.

Fixed

• Compressed Codex logs are read. Current Codex compresses older session logs to rollout-*.jsonl.zst and keeps only the newest ones as plain .jsonl. Sipcode skipped the compressed ones, so it saw almost none of a Codex user's history: forecast said it needed more days, limits started again at 0%, and stats, why, today, trend, impact and regress missed most sessions. Sipcode now reads them exactly like plain logs, on Node.js 22.15 or newer (zstd ships in Node's own zlib, so there is no new dependency and still no network). A session with both a plain and a compressed log is read once, from the plain log. A damaged or cut-off compressed log is counted as unreadable, never a crash.
• Older Node.js: a clear warning. On a Node.js without zstd the compressed logs are still counted and reported (see "Logs Sipcode could not read" below). JSON output is unchanged; with --json the note "N compressed Codex log(s) (.jsonl.zst) skipped: reading them needs Node.js 22.15 or newer (you have vX)." goes to stderr.
• regress sandbox warning with --agent. The "could not be read" WARNING told the user to allow read access to both log folders, and with --agent codex it named an empty one ("allow read access to and ..."). It now names only the folders the run read.
• regress --bug-report with nothing to report says so. With no finding, the flag used to print only the plain result, which Codex read as "no output". It now ends with "No bug report: there is no finding to report." The answer was and is always on stdout.

Changed

• Logs Sipcode could not read are impossible to miss. When some log files cannot be read (compressed Codex logs on a Node.js older than 22.15, a damaged or unreadable file, a folder that cannot be listed), stats, today, forecast, trend, impact, why, receipt, drift, limits and regress now start their terminal output with one line per reason, for example "warning: 3 Codex log file(s) could not be read (compressed .jsonl.zst logs need Node.js 22.15 or newer; you have v22.11.0). Numbers below leave them out." It used to be a small note, often the last line and on stderr, and why, receipt and drift did not mention these logs at all. With --json the notes stay on stderr and the JSON is unchanged (stats keeps its warnings array). When every log is read, the output is byte for byte what it was. The four Codex plugin skills say this warning first in their answer.
• Codex plugin: regress checks the Codex logs by default. A first regress over large Claude Code logs can take a minute or more (49 seconds on 5.4 GB on the maintainer's machine; about 1 second once the usage cache is warm), Codex stopped the command after about 10 seconds with empty output, and a sandbox that cannot write ~/.sipcode never warms the cache. When the user asks about Codex or names no tool, the regress skill now runs sipcode regress --bug-report --agent codex (Codex logs only, about 1 second). It reads the Claude Code logs only when the user asks about Claude Code, says first that the run can take a minute or more, and gives the command a timeout of at least 3 minutes. Empty output is never reported as a result.
• README: Node.js 22.15 or newer is needed to read compressed Codex logs.

How it works

• One helper reads every session log (readLogText in src/lib/fs.ts); Codex discovery decodes only line 1 of a compressed log, by streaming up to the first newline. The usage cache keeps working (path, size, mtime).
• src/lib/zstd.ts walks the zstd frames before decoding, because Node decodes only the first frame of a multi-frame file and returns partial text with no error for a truncated one. Each frame's decoded size is checked against its header.
• The unread-logs warning has one source (unreadWarnings in src/modules/agents/loadSessions.ts). Period commands collect it per section (SectionOutput.unread) and runSections prints every section's warnings above the banner and the section headers. why / receipt / drift wrap their output in UnreadOutput (src/modules/agents/latest.ts), which prints discovery's counts plus the logs the pick or drift's baseline could not read before the first line the command writes. limits warns on the "no readings" path too; regress puts the warnings above its report.

Notes

• scripts/verify-counts.mjs (the independent counter) reads .jsonl.zst too, decoding it with node:zlib itself, so it cannot agree with a Sipcode that skips them.
• Test count: 2,225 to 2,409 (npx vitest run, which includes the 20 e2e tests).

Verification

• npm run lint, npm run build: pass. npx vitest run: 2,409 passed, twice in a row. npm run test:e2e: 20 passed.
• New tests: a Codex home laid out plain, mixed, all compressed, and with both forms gives byte-identical output for stats, today, forecast, limits, why (newest, --session, a resumed log), receipt, drift and regress, JSON and terminal. Per reader: discovery line 1, full parse, key scan of logs outside a window, usage cache cold and warm, damaged and truncated files, and a Node.js without zstd (injected, not by removing zlib). Reverting any one reader to a plain read fails these tests.
• Unread-logs warning: a golden file recorded before the change proves every command's output (stats, today, forecast, trend, impact, why, receipt, drift, limits, regress; terminal and JSON; both tools and each alone) is byte-identical when every log is read. The same 58 cases run with two compressed logs skipped (Node.js without zstd) and with two damaged ones: in the terminal the warning is line 1 with the exact count, followed by the unchanged output, and no stderr note; with --json stdout is the same JSON and the note is on stderr. Also covered: limits with every log skipped (warning before "no readings"), why --agent codex with every log skipped (warning before the error), a skipped and an unreadable log in one run, the regress formatter, runSections and UnreadOutput.
• Codex plugin: the regress skill's first command is sipcode regress --bug-report --agent codex; the Claude Code commands are listed with the "only when the user explicitly asks" rule; every skill says the warning first. Every command a skill names exists in src and dist.
• Real data, unread-logs warning (a copy of the maintainer's Codex logs in a scratch folder, scratch HOME; the real folders untouched): with one extra damaged .jsonl.zst in the copy, stats --agent codex starts with "warning: 1 Codex log file(s) could not be read (unreadable or damaged). Numbers below leave them out.", and today, forecast, trend, impact, why, receipt, drift, limits and regress do the same; stats --agent codex --json keeps the JSON (its warnings array has the E003 entry). Without the extra file there is no warning line.
• Real data, regress --bug-report --agent codex (same copy: 17 Codex logs, 15 compressed; fresh scratch HOME): 1.8 s cold, 1.1 s warm, exit 0; "Codex: 951 requests in 17 sessions, 2026-03-28 to 2026-10-15, versions 0.118.0 to 0.162.0, 4 version/model groups with enough limit data", then "No regressions detected in your logs." and "No bug report: there is no finding to report."
• Real data (frozen copy of the maintainer's logs, scratch HOME): verify-counts all checks match, exit 0. Codex: 17 logs (15 compressed), 951 requests, 89,451,779 tokens, $56.66; every log equals Codex's own final total; limits again shows the possible cut around Apr 13 (5.1x); forecast has 14 days of history.

🤖 Generated with Claude Code