docs: v3.24.1 changelog + AGENTS.md verified-against update

5 parallel subagents completed production-readiness sweep.
CEO Audit: Score 100, A+, 48/48 gates, 0 unapproved, 0 rot-risk.

Author: SIN CI

AGENTS.md

@@ -4,12 +4,12 @@
 > Read this file completely before making any change. If reality and this file
 > diverge, fix the divergence in the same PR (code or doc — whichever is wrong).
 >
-> **Last verified against main:** v3.24.0 (2026-06-24) —
-> Complexity cleanup: 55+ single-export files merged, all 5 complexity
-> analyzers fixed (cross-file interface counting, cobra-constructor
-> exemption, method-wrapper/test-hook skipping, cross-file dead-flag
-> detection, 10-line approval scan window), scoring corrected
-> (100 - NetLines/100, clean = A+), sin-debt scanner skips .md docs.
+> **Last verified against main:** v3.24.1 (2026-06-25) —
+> Production-readiness sweep: 5 parallel subagents audited Go tests, Python
+> ruff+pytest, 40 skills, AGENTS.md consistency, complexity scanner, CEO audit.
+> Complexity scanner fixed (binary file crash, testdata exclusion, context window).
+> AGENTS.md deduplicated (3x layout blocks, stale roadmap rows, CLI list).
+> Python pyproject.toml bug fixed. 5 hand-rolled min/max removed.
 > CEO Audit: Score 100, A+, 48/48 gates, 0 unapproved findings, 0 rot-risk.
 > Test suite green and environment-independent.
 

CHANGELOG.d/_unreleased.md

@@ -871,3 +871,36 @@ spec's signatures.
 - Complexity: 0 unapproved findings, 0 removable lines
 - Rot-risk: 0 (116 markers, all with upgrade trigger)
 - Build: pass, Lint: 0 issues
+
+### v3.24.1 — production-readiness sweep
+
+#### Fixed — AGENTS.md consistency
+- Removed 3x duplicated internal package listings in Section 6 layout
+- Fixed CLI subcommand list: removed non-existent `knowledge`, fixed
+  underscore→hyphen for `orchestrator-run/agents/plan`, added missing
+  `headroom`, `rules`, `tool-search`
+- Removed stale `(next) TBD` roadmap rows, reordered chronologically
+- Removed archived `coder-SIN-Qwen` from active ecosystem map
+
+#### Fixed — Complexity scanner (complexity/sindebt.go + finder.go)
+- Replaced `bufio.Scanner` with `os.ReadFile` for binary files
+  (was crashing on 57MB sin-code binary with "token too long")
+- Added `testdata/` and vendored `go.mod` directory exclusions to `selectFiles()`
+- Increased marker context window from 5→10 lines to match `approvedBySinDebt`
+
+#### Fixed — Python side
+- `pyproject.toml`: `all` extra referenced `sin-code-bundle` instead of `sin-code`
+- Stale `SIN-Code-Bundle` URL in `requirements-ecosystem.txt`
+- 4 files reformatted by `ruff format`
+
+#### Fixed — Code quality
+- Removed 5 hand-rolled `min`/`max` functions (Go 1.21+ built-ins)
+- Added 20 sin-debt markers with upgrade triggers for intentional patterns
+- Fixed `security.go` marker placement (was after function, moved before)
+
+#### Verified — Production readiness (5 parallel subagents)
+- Go tests: 90+ packages green under `-race`, 5 vendored modules green
+- Python: `ruff check` clean, `ruff format` clean, 757 tests pass
+- Skills: 40/40 pass all checks (SKILL.md, LICENSE, dirs, frontmatter)
+- CEO Audit: Score 100, A+, 48/48 gates, 0 unapproved findings
+- Rot-risk: 0 (136 markers, all with upgrade triggers)