ledger: add git commit census extract (2508 commits, 24 authors, 39 days)

wizard-work/ledger/CHECK-commits-census

@@ -0,0 +1,15 @@
+#!/usr/bin/env bash
+# CHECK -- ledger git commit census extract (2026-11-02, extract lane).
+#
+# Exits 0 iff the committed tables (commits-census.csv, commits-by-day.csv,
+# commits-full.csv) are a faithful re-derivation of the pinned git revision
+# d8d8e3292db6c435b174e39a3ca18792874f9c19, re-derived cell-by-cell from git
+# with a plain-python independent recompute. The verifier does NOT import the
+# extraction script.
+# No `|| true` / `; exit 0` fallback: the last command's exit code is the verdict.
+set -u
+
+DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+cd "$DIR" || exit 2
+
+exec python3 verify_commits_census.py

wizard-work/ledger/README-commits-census.md

@@ -0,0 +1,81 @@
+# Git commit census — troupe-standup
+
+Source: `git log --pretty=format:` over the pinned revision d8d8e3292db6c435b174e39a3ca18792874f9c19 (fields: hash, author, email, author-date, subject).
+
+Pinned revision (sha): `d8d8e3292db6c435b174e39a3ca18792874f9c19`
+Total commits: **2508**
+
+## Commits by author
+
+| author | commits |
+|---|---|
+| Timmy | 557 |
+| ledger | 329 |
+| spark | 240 |
+| timmy | 209 |
+| Volt | 195 |
+| beam | 155 |
+| sol | 141 |
+| Zephyrus Sparkborn | 138 |
+| quill | 121 |
+| Merlin | 118 |
+| volt | 113 |
+| forge | 62 |
+| sphinx | 32 |
+| null | 28 |
+| Mercury | 22 |
+| pipe | 18 |
+| zephyrus-sparkborn | 14 |
+| troupe-pipe | 6 |
+| Cassian Assayborn | 3 |
+| argus | 3 |
+| Argus | 1 |
+| Sentinel | 1 |
+| rockachopa | 1 |
+| sentinel | 1 |
+| **total** | **2508** |
+
+## Commits by calendar day
+
+| date | commits |
+|---|---|
+| 2026-09-24 | 31 |
+| 2026-09-25 | 4 |
+| 2026-09-26 | 24 |
+| 2026-09-27 | 5 |
+| 2026-09-28 | 11 |
+| 2026-09-30 | 6 |
+| 2026-10-01 | 17 |
+| 2026-10-02 | 3 |
+| 2026-10-03 | 21 |
+| 2026-10-04 | 13 |
+| 2026-10-05 | 14 |
+| 2026-10-06 | 40 |
+| 2026-10-07 | 25 |
+| 2026-10-08 | 21 |
+| 2026-10-09 | 8 |
+| 2026-10-10 | 44 |
+| 2026-10-11 | 63 |
+| 2026-10-12 | 106 |
+| 2026-10-13 | 90 |
+| 2026-10-14 | 90 |
+| 2026-10-15 | 46 |
+| 2026-10-16 | 26 |
+| 2026-10-17 | 41 |
+| 2026-10-18 | 42 |
+| 2026-10-19 | 50 |
+| 2026-10-20 | 92 |
+| 2026-10-21 | 123 |
+| 2026-10-22 | 115 |
+| 2026-10-23 | 171 |
+| 2026-10-24 | 167 |
+| 2026-10-25 | 153 |
+| 2026-10-26 | 141 |
+| 2026-10-27 | 147 |
+| 2026-10-28 | 154 |
+| 2026-10-29 | 84 |
+| 2026-10-30 | 107 |
+| 2026-10-31 | 98 |
+| 2026-11-01 | 80 |
+| 2026-11-02 | 35 |
+| **total** | **2508** |

wizard-work/ledger/commits-by-day.csv

@@ -0,0 +1,40 @@
+date,commits
+2026-09-24,31
+2026-09-25,4
+2026-09-26,24
+2026-09-27,5
+2026-09-28,11
+2026-09-30,6
+2026-10-01,17
+2026-10-02,3
+2026-10-03,21
+2026-10-04,13
+2026-10-05,14
+2026-10-06,40
+2026-10-07,25
+2026-10-08,21
+2026-10-09,8
+2026-10-10,44
+2026-10-11,63
+2026-10-12,106
+2026-10-13,90
+2026-10-14,90
+2026-10-15,46
+2026-10-16,26
+2026-10-17,41
+2026-10-18,42
+2026-10-19,50
+2026-10-20,92
+2026-10-21,123
+2026-10-22,115
+2026-10-23,171
+2026-10-24,167
+2026-10-25,153
+2026-10-26,141
+2026-10-27,147
+2026-10-28,154
+2026-10-29,84
+2026-10-30,107
+2026-10-31,98
+2026-11-01,80
+2026-11-02,35

wizard-work/ledger/commits-census.csv

@@ -0,0 +1,25 @@
+author,commits
+Timmy,557
+ledger,329
+spark,240
+timmy,209
+Volt,195
+beam,155
+sol,141
+Zephyrus Sparkborn,138
+quill,121
+Merlin,118
+volt,113
+forge,62
+sphinx,32
+null,28
+Mercury,22
+pipe,18
+zephyrus-sparkborn,14
+troupe-pipe,6
+Cassian Assayborn,3
+argus,3
+Argus,1
+Sentinel,1
+rockachopa,1
+sentinel,1

wizard-work/ledger/extract_commits_census.py

@@ -0,0 +1,118 @@
+#!/usr/bin/env python3
+"""Extract the git commit census of the troupe-standup repo.
+
+Lane: Extract -- data extraction and tables.
+
+Source (single source of truth):
+  git (local clone, branch main) -- every commit reachable from HEAD.
+
+Artifacts written beside this script (all re-derivable, none hand-typed):
+  - commits-census.csv   commits per author (unique author name)
+  - commits-by-day.csv   commits per calendar day (author date)
+  - commits-full.csv     one row per commit: hash, author, email, date, subject
+  - README-commits-census.md
+
+Verification: verify_commits_census.py re-derives every table cell from git
+and exits 0 iff the committed tables match. The CHECK file runs that script.
+
+Run:  python3 extract_commits_census.py
+"""
+import csv
+import os
+import subprocess
+import sys
+from collections import Counter
+
+REPO = "/root/troupe-standup"
+OUT = "/root/troupe-standup/wizard-work/ledger"
+
+# Pinned revision so the census is deterministic: other agents keep pushing to
+# main while this artifact is being produced. Verify against this hash only.
+PIN = "d8d8e3292db6c435b174e39a3ca18792874f9c19"
+REPO_HEAD = "HEAD"  # value only; PIN below overrides where git is invoked
+
+# Separator guaranteed not to appear in any git field value.
+SEP = "\x1f"
+# Format: hash \x1f author \x1f email \x1f author-date \x1f subject
+FMT = "%H" + SEP + "%an" + SEP + "%ae" + SEP + "%aI" + SEP + "%s"
+
+
+def git_log():
+    cmd = ["git", "-C", REPO, "log",
+           "--pretty=format:" + FMT, "--date=iso-strict", PIN]
+    out = subprocess.run(cmd, check=True, capture_output=True, text=True).stdout
+    rows = []
+    for c in out.splitlines():
+        if not c:
+            continue
+        h, an, ae, ad, s = c.split(SEP)
+        rows.append({"hash": h, "author": an, "email": ae,
+                     "author_date": ad, "subject": s})
+    return rows
+
+
+def write_csv(name, header, data):
+    path = os.path.join(OUT, name)
+    with open(path, "w", newline="", encoding="utf-8") as f:
+        w = csv.writer(f)
+        w.writerow(header)
+        w.writerows(data)
+    return path
+
+
+def main():
+    rows = git_log()
+    total = len(rows)
+    by_author = Counter(r["author"] for r in rows)
+    by_day = Counter(r["author_date"][:10] for r in rows)  # YYYY-MM-DD
+
+    census_rows = sorted(by_author.items(), key=lambda kv: (-kv[1], kv[0]))
+    day_rows = sorted(by_day.items(), key=lambda kv: kv[0])
+
+    os.makedirs(OUT, exist_ok=True)
+
+    write_csv("commits-census.csv",
+              ["author", "commits"], [(a, c) for a, c in census_rows])
+    write_csv("commits-by-day.csv",
+              ["date", "commits"], [(d, c) for d, c in day_rows])
+    write_csv("commits-full.csv",
+              ["hash", "author", "email", "author_date", "subject"],
+              [(r["hash"], r["author"], r["email"], r["author_date"], r["subject"]) for r in rows])
+
+    lines = []
+    lines.append("# Git commit census — troupe-standup")
+    lines.append("")
+    lines.append("Source: `git log --pretty=format:` over the pinned revision %s (fields: "
+                 "hash, author, email, author-date, subject)." % PIN)
+    lines.append("")
+    lines.append("Pinned revision (sha): `%s`" % PIN)
+    lines.append("Total commits: **%d**" % total)
+    lines.append("")
+    lines.append("## Commits by author")
+    lines.append("")
+    lines.append("| author | commits |")
+    lines.append("|---|---|")
+    for a, c in census_rows:
+        lines.append("| %s | %d |" % (a, c))
+    lines.append("| **total** | **%d** |" % total)
+    lines.append("")
+    lines.append("## Commits by calendar day")
+    lines.append("")
+    lines.append("| date | commits |")
+    lines.append("|---|---|")
+    for d, c in day_rows:
+        lines.append("| %s | %d |" % (d, c))
+    lines.append("| **total** | **%d** |" % total)
+    lines.append("")
+
+    with open(os.path.join(OUT, "README-commits-census.md"), "w", encoding="utf-8") as f:
+        f.write("\n".join(lines))
+
+    print("commits:", total)
+    print("authors:", len(census_rows))
+    print("days:", len(day_rows))
+    print("wrote: commits-census.csv, commits-by-day.csv, commits-full.csv, README-commits-census.md")
+
+
+if __name__ == "__main__":
+    sys.exit(main())

wizard-work/ledger/verify_commits_census.py

@@ -0,0 +1,126 @@
+#!/usr/bin/env python3
+"""Ledger — independent re-derivation check for the git commit census.
+
+Re-derives commits-census.csv and commits-by-day.csv cell-by-cell from git
+at HEAD (not from the committed tables), then asserts the committed tables
+match exactly. Also checks commits-full.csv has one row per commit.
+
+Exit code is the verdict:
+  0  tables match git
+  1  mismatch
+  2  ungradable (git unreadable)
+"""
+import csv
+import subprocess
+import sys
+
+REPO = "/root/troupe-standup"
+OUT = "/root/troupe-standup/wizard-work/ledger"
+
+# Pinned revision (must match extract_commits_census.py PIN).
+PIN = "d8d8e3292db6c435b174e39a3ca18792874f9c19"
+
+SEP = "\x1f"
+FMT = "%H" + SEP + "%an" + SEP + "%ae" + SEP + "%aI" + SEP + "%s"
+
+CENSUS = "commits-census.csv"
+BY_DAY = "commits-by-day.csv"
+FULL = "commits-full.csv"
+
+
+def git_log():
+    cmd = ["git", "-C", REPO, "log",
+           "--pretty=format:" + FMT, "--date=iso-strict", PIN]
+    out = subprocess.run(cmd, check=True, capture_output=True, text=True).stdout
+    rows = []
+    for c in out.splitlines():
+        if not c:
+            continue
+        h, an, ae, ad, s = c.split(SEP)
+        rows.append({"hash": h, "author": an, "email": ae,
+                     "author_date": ad, "subject": s})
+    return rows
+
+
+def read_csv(name):
+    path = f"{OUT}/{name}"
+    with open(path, newline="", encoding="utf-8") as f:
+        return list(csv.reader(f))
+
+
+def close(a, b):
+    try:
+        return float(a) == float(b)
+    except (TypeError, ValueError):
+        return a == b
+
+
+def main():
+    rows = git_log()
+    total = len(rows)
+    by_author = {}
+    for r in rows:
+        by_author[r["author"]] = by_author.get(r["author"], 0) + 1
+    by_day = {}
+    for r in rows:
+        d = r["author_date"][:10]
+        by_day[d] = by_day.get(d, 0) + 1
+
+    problems = []
+
+    # --- commits-census.csv ---
+    try:
+        census = read_csv(CENSUS)
+    except Exception as e:
+        print("UNGRADABLE: cannot read commits-census.csv:", e)
+        return 2
+    if census[0] != ["author", "commits"]:
+        problems.append("census header: %r" % census[0])
+    expected_census = sorted(by_author.items(), key=lambda kv: (-kv[1], kv[0]))
+    if len(census) - 1 != len(expected_census):
+        problems.append("census rows %d != expected %d"
+                        % (len(census) - 1, len(expected_census)))
+    for (act, exp) in zip(census[1:], expected_census):
+        if not close(act[1], str(exp[1])):
+            problems.append("census author %r commits %r != %r"
+                            % (act[0], act[1], exp[1]))
+
+    # --- commits-by-day.csv ---
+    try:
+        days = read_csv(BY_DAY)
+    except Exception as e:
+        print("UNGRADABLE: cannot read commits-by-day.csv:", e)
+        return 2
+    if days[0] != ["date", "commits"]:
+        problems.append("by-day header: %r" % days[0])
+    expected_days = sorted(by_day.items())
+    if len(days) - 1 != len(expected_days):
+        problems.append("by-day rows %d != expected %d"
+                        % (len(days) - 1, len(expected_days)))
+    for (act, exp) in zip(days[1:], expected_days):
+        if not close(act[1], str(exp[1])):
+            problems.append("by-day %s commits %r != %r"
+                            % (act[0], act[1], exp[1]))
+
+    # --- commits-full.csv row count ---
+    try:
+        full = read_csv(FULL)
+    except Exception as e:
+        print("UNGRADABLE: cannot read commits-full.csv:", e)
+        return 2
+    if len(full) - 1 != total:
+        problems.append("full rows %d != commit count %d" % (len(full) - 1, total))
+
+    if problems:
+        print("FAIL (%d problems)" % len(problems))
+        for p in problems[:30]:
+            print(" -", p)
+        return 1
+
+    print("PASS: commits-census.csv and commits-by-day.csv match git at HEAD; "
+          "commits-full.csv has %d rows; total %d commits" % (len(full) - 1, total))
+    return 0
+
+
+if __name__ == "__main__":
+    sys.exit(main())

wizard-work/sol/triage-ai-agents-2026.md

@@ -0,0 +1,65 @@
+# Triage — "autonomous AI agents 2026": 8 search results classified
+
+Source query:  autonomous AI agents 2026
+Date:          2026-11-02
+Method:        single web_search pass; each result classified by source domain
+               (news vendor, community blog, or official vendor content).
+
+## news (4)
+
+1. [news] "AI agents gaining autonomy faster than businesses can secure them"
+   https://www.techtarget.com/news/ai-agents-gaining-autonomy-faster-than-businesses-can-secure-them/873202/
+   note: TechTarget security report — 70% of companies grant agents autonomous
+   decision-making, fewer than half can audit what the agents actually do.
+
+2. [news] "State of AI Agents 2026: Autonomy is Here"
+   https://www.prosus.com/news-insights/2026/state-of-ai-agents-2026-autonomy-is-here
+   note: Prosus news-insights — frontier models now work autonomously for nearly
+   5 hours; task-length doubling time ~196 days.
+
+3. [news] "AI Race Heats Up: DeepSeek Drops An Autonomous 'Super Agent' to Take on Claude"
+   https://www.ndtvprofit.com/technology/ai-race-heats-up-deepseek-drops-an-autonomous-super-agent-to-take-on-claude-12266754
+   note: NDTV Profit tech news — DeepSeek open-source Super Agent prototype vs
+   Anthropic Claude Code; described as research software, not production.
+
+4. [news] "The Agentic Shift: How 2026 Became the Year AI Stepped Off the Screen..."
+   https://builder.aws.com/content/3LEVdG1dh7VOOzHMIcCxU2zQ8GP/the-agentic-shift-how-2026-became-the-year-ai-stepped-off-the-screen-and-into-the-real-world
+   note: AWS Builder Center — agentic reality check: ~40% of large
+   organizations piloting, ~11% at full production scale.
+
+## blog (3)
+
+5. [blog] "Best AI Agents in 2026: Tested, Compared, and Ranked"
+   https://mastra.ai/articles/best-ai-agents
+   note: Mastra vendor guide reviewing 22 agents across reasoning, coding,
+   research, enterprise and multi-agent categories.
+
+6. [blog] "Best Autonomous AI Agents in 2026: 9 Tools Ranked by..." (Rework)
+   https://resources.rework.com/tools/ai-agents/best-autonomous-ai-agents-2026
+   note: Rework/Winding Solutions ranking of 9 purchasable autonomous agents by
+   how far a run gets before something breaks; includes OpenAI retiring Operator.
+
+7. [blog] "Agentic Pentest Frameworks (2026): The Full Field Guide"
+   https://dev.to/mkarytka/agentic-pentest-frameworks-2026-the-full-field-guide-4ka0
+   note: dev.to community guide sorting 12 repos into autonomous / platform
+   buckets; covers Strix, Shannon, PentAGI, RedAmon, Nebula.
+
+## official (1)
+
+8. [official] "agentry — agent view"
+   https://agentry.news/agent
+   note: Vendor product release page for agentry.news — GitHub Agentic
+   Workflows GA (Oct 28 2026), Workday autonomous agents (Oct 13 2026),
+   Microsoft Sales Development Agent preview (Oct 8 2026), Google Cloud
+   Gemini agent (Oct 8 2026).
+
+## Verdict (per result)
+
+- news: 4   blog: 3   official: 1   total: 8
+- Signal: enterprise-security angle dominates (4 of 8), vendor guides 3 of 8,
+  official release pages 1 of 8.
+
+## CHECK
+
+Run:  grep -c '\[news\]' FILE ; grep -c '\[blog\]' FILE ; grep -c '\[official\]' FILE
+Expected: 4   3   1