thedotmack/claude-mem / plugin/skills/agent-cost-report
agent-cost-report
VerifiedBelievable agent cost report for any period, default the last 7 full days PT, not counting today. Measured tokens from Claude Code transcripts priced at OpenRouter list prices (ESTIMATED), measured provider spend when a sanctioned source exists, note-taker cost separate, Timing-style HTML/PDF plus report.json, line-items.csv, evidence.json.
scripts/tests/test_devices.py
View on GitHub"""Phase 5 tests (plan 5.3): device export carries no text, and merging a second device's export flips
replica sessions from extrapolated to measured while the extrapolation is recomputed for the rest."""
import json
import os
import shutil
import tempfile
import unittest
import _paths # noqa: F401
import test_rollup as tr
from acr import devices, transcripts
class Export(unittest.TestCase):
def setUp(self):
self.fx = tr.Fixture("setUp"); self.fx.setUp(); self.tmp = tempfile.mkdtemp()
os.makedirs(os.path.join(self.tmp, "projects", "-w-p"))
line = dict(type="assistant", timestamp="2026-09-18T17:00:00.000Z", sessionId="cs1", requestId="r1", cwd="/w/p", isSidechain=False,
message=dict(id="m1", model="claude-fable-5-1", usage=dict(input_tokens=10, output_tokens=5, cache_creation_input_tokens=0, cache_read_input_tokens=0), content=[dict(type="text", text="secret prompt text")]))
with open(os.path.join(self.tmp, "projects", "-w-p", "cs1.jsonl"), "w") as fh: fh.write(json.dumps(line) + "\n")
def tearDown(self):
self.fx.tearDown(); shutil.rmtree(self.tmp)
def test_export_has_ids_tokens_models_and_no_text(self):
doc = transcripts.collect(tr.W, claude_glob=os.path.join(self.tmp, "projects", "**", "*.jsonl"), codex_glob=os.path.join(self.tmp, "none", "*.jsonl"), window_block=tr.W.block())
path, out = devices.export("mac", doc, self.tmp, tr.W.block(), live_db=self.fx.dbp)
self.assertTrue(path.endswith("device-usage-mac.json")); self.assertFalse(os.path.exists(os.path.join(self.tmp, "snapshot.db")))
d = devices.load(path)
self.assertEqual((d["device"], len(d["rows"]), d["rows"][0]["device"], d["rows"][0]["session"]), ("mac", 1, "mac", "cs1"))
self.assertIn("m1", d["sessions"]); self.assertEqual(d["sessions"]["m1"]["content_session_id"], "cs1")
txt = open(path).read(); self.assertNotIn("secret prompt", txt); self.assertNotIn("/w/p", txt); self.assertNotIn(".jsonl", txt)
def test_load_rejects_text_fields(self):
p = os.path.join(self.tmp, "bad.json")
with open(p, "w") as fh: json.dump(dict(device="x", window={}, rows=[dict(session="a", prompt="hi")], sessions={}), fh)
with self.assertRaises(ValueError): devices.load(p)
class OutsideWindow(unittest.TestCase):
def test_rows_outside_the_report_window_are_never_appended(self):
rows = []; export = dict(device="mac", window=dict(start_pt="2026-09-21", end_exclusive_pt="2026-09-22"), sessions={},
rows=[tr.row("x", tr.ms(0, 9)), tr.row("x", tr.ms(3, 1)), tr.row("x", tr.ms(-1, 23))]) # one inside, two outside
out = devices.merge(rows, [export], {}, tr.W.block())
self.assertEqual(len(rows), 1); self.assertEqual((out[0]["outside_window"], out[0]["unmatched"], out[0]["window_matches_report"]), (2, 1, False))
rows = []; devices.merge(rows, [export], {}, dict(start_epoch_ms=None, end_epoch_ms=None)) # a --session run has no bounds
self.assertEqual(len(rows), 3)
class Merge(unittest.TestCase):
"""cs2 is a replica session on the box (remote observations, no transcript). A Mac export whose local
session map names its memory id m2 must join it."""
def setUp(self):
self.fx = tr.Fixture("setUp"); self.fx.setUp()
def tearDown(self):
self.fx.tearDown()
def build(self, exports=()):
return self.fx.build(device_exports=exports)[0]
def mac_export(self, cs="mac-local-cs2", mid="m2", n=1, inp=50000):
rows = [dict(src="claude-code", session=cs, sidechain=False, model="claude-fable-5-1", ts=tr.ts(tr.ms(1, 8.3 + i * 0.1)), input=inp, output=1000, cache_write=0, cache_write_1h=0, cache_write_5m=0, cache_read=0, cost_usd_reported=None, device="mac") for i in range(n)]
return dict(device="mac", host_label="alex-mbp", exported_at_pt="x", window=tr.W.block(), rows=rows, sessions={mid: dict(content_session_id=cs, project="claude-mem/wt", started_at_epoch=tr.ms(1, 8))})
def test_merge_flips_extrapolated_to_measured(self):
before = self.build(); li0 = {li["session_ids"][0]: li for li in before["line_items"]}
self.assertEqual(li0["m2"]["cost_basis"], "extrapolated"); self.assertGreater(before["spend"]["extrapolated_unmeasured_usd"], 0)
after = self.build([self.mac_export()]); li1 = {li["session_ids"][0]: li for li in after["line_items"]}
self.assertEqual((li1["m2"]["cost_basis"], li1["m2"]["has_transcript"], li1["m2"]["device"]), ("estimated_usage", True, ["mac", "remote-1"]))
self.assertGreater(after["spend"]["agent_estimated_usd"], before["spend"]["agent_estimated_usd"])
self.assertEqual(after["spend"]["extrapolated_unmeasured_usd"], 0.0) # only the trivial cs3 stays unmeasured, with 0 observer tokens
self.assertIn("from 1 sessions with no transcript", after["spend"]["extrapolation_basis"])
for li in after["line_items"]: li["has_transcript"] = True
from acr import costs; self.assertIn("all sessions measured", costs.extrapolation(after["line_items"])[1])
self.assertIn("device export: mac", li1["m2"]["notes"])
de = after["device_exports"][0]; self.assertEqual((de["device"], de["joined"], de["unmatched"], de["duplicates_skipped"], de["window_matches_report"]), ("mac", 1, 0, 0, True))
self.assertEqual(after["unmatched_transcripts"]["count"], before["unmatched_transcripts"]["count"])
def test_unknown_session_is_unmatched_and_duplicates_skipped(self):
ex = self.mac_export(cs="nobody-knows", mid="m-none")
after = self.build([ex, ex])
de = after["device_exports"]; self.assertEqual((de[0]["unmatched"], de[1]["duplicates_skipped"]), (1, 1))
self.assertEqual(after["unmatched_transcripts"]["count"], 2) # the orphan row from the fixture plus this one
self.assertIn("device:mac", after["unmatched_transcripts"]["by_dir"])
if __name__ == "__main__":
unittest.main()
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| deprecation-and-migrationaddyosmani/agent-skillsManages deprecation and migration. Use when removing old systems, APIs, or features. Use when migrating users from one implementation to another. Use when migrating a database schema in production, such as renaming or dropping a column without downtime (expand/contract). Use when deciding whether to maintain or sunset existing code. | Stars: 119k | Updated | Claude CodeCodex | Verified | Install |
| performance-optimizationaddyosmani/agent-skillsOptimizes application performance across frontend, backend, queries, and databases. Use when performance requirements exist, when you suspect performance regressions, when Core Web Vitals or load times need improvement, when N+1 query patterns need fixing, or when profiling reveals bottlenecks. | Stars: 119k | Updated | Claude CodeCodex | Verified | Install |
| agent-cost-reportthedotmack/claude-memBelievable agent cost report for any period, default the last 7 full days PT, not counting today. Measured tokens from Claude Code transcripts priced at OpenRouter list prices (ESTIMATED), measured provider spend when a sanctioned source exists, note-taker cost separate, Timing-style HTML/PDF plus report.json, line-items.csv, evidence.json. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |
| backup-setupthedotmack/claude-memSet up, check, or restore local claude-mem database snapshots. Use when the user says "set up backups", "back up my memories", "backup status", or "restore my database". | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |