IPOT's investor filter splits each broker row by client domicile, but its
footer then gives only that side's buying (BBCA 2026-10-16: foreign 1.7 T,
domestic 4.9 T of 6.6 T). The foreign net gives the selling on all boards:
foreign 1.7 + 2.3 = 4.0 T (IDX: about 3.95 T regular), domestic
4.9 - 2.3 = 2.6 T.
- BrokerSummary gains buy_value and sell_value: the total for all
investors, the footer and the derived selling for one side (selling is
null with a board filter, where the net covers other boards).
- Bandar weighs the top N buyers' share of the value bought against the
top N sellers' share of the value sold, so `--investor foreign|domestic`
reads that side alone; for all investors the readings are unchanged.
One side on one board is refused, as there is no selling to weigh.
- Tables print what the side bought and sold; fixtures for foreign and
domestic on all boards were recorded from IPOT.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01V3syn2fN2LNo5pvTSqc6eh
IDX's Value is regular-board turnover, so NonRegularValue / Value gave
non-regular over regular (BBRI 2026-10-14: 50.91%) instead of the share of
all trading (33.74%). non_regular_bps now divides by regular plus
non-regular value, with a regression test on that fixture. The docs now say
volume, value and the foreign figures are the regular board: IPOT's
regular total matches Value, and in a week of every stock no foreign figure
exceeded regular volume, even with non-regular volume above twice it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01V3syn2fN2LNo5pvTSqc6eh
`idx stocks flow SYMBOL` shows what foreign investors bought and sold of a
stock, session by session, from IDX's own daily trading summary: foreign
buy, sell and net volume, net value at the session's average price, and
the share of value traded outside the regular board, with range totals.
Latest 10 sessions by default (--days, at most 60) or --from/--to. The MCP
server exposes it as get_foreign_flow.
- api/idx: pure parser for GetStockSummary (rows dated another session
are dropped) and a provider over curl-impersonate's profile fallback
with a 30 s limit; Cloudflare passes only some fingerprints.
- flow: each session is cached whole by date, so other stocks' flows for
the same sessions need no request. Weekends are never requested.
- Idx::idx_code is shared by broker and flow for the .JK-only check.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01V3syn2fN2LNo5pvTSqc6eh
- Basis points truncate toward zero instead of flooring, so a reading of
-4.9952% stays neutral like its mirror image and -14.9952% stays
distribution; tests cover both signs at both thresholds through
bandar_tier. Goldens move by the rounding (BBCA top 5: -2426 -> -2425).
- With one investor side selected the summary has no total to weigh
brokers against, so bandar and analyze_bandar reject it (INVALIDINPUT)
instead of reading every tier as missing.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01V3syn2fN2LNo5pvTSqc6eh
`idx stocks bandar SYMBOL...` reads the broker summary: for the top 1, 3,
5 and 10 brokers per side, buy value minus sell value in basis points of
value traded, with a signal from big accumulation to big distribution on
the headline tier (--top, default 5). Several symbols give a screen ranked
by that figure (at most 100, one request each), which pipes from
`stocks screen`. The MCP server exposes it as analyze_bandar (20 symbols).
- analysis/bandar.rs is pure and versioned (`top_concentration_v1`):
integer basis points, no figure missing counted as zero, buy/sell VWAP
from exact averages, and the brokers on both sides listed because their
gross figures overstate their direction.
- Thresholds (±5%/±15%) follow fzn0x/idx (MIT); docs call them heuristics.
- Invalid input fails a screen at once; other errors fail one symbol and
are warned about on stderr.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01V3syn2fN2LNo5pvTSqc6eh
`idx stocks broker SYMBOL` shows the top 10 buyers and sellers of a stock
by lots for the latest session with trades, one `--date`, or a
`--from`/`--to` range, filtered by `--investor` and `--board`. The MCP
server exposes the same as get_broker_summary.
- Range and filter parsing lives in idx-core (BrokerRange::from_args,
FromStr for the filters), so both frontends accept the same input.
- Without dates, the latest session is found by walking back over
weekdays; offline, uncached days are skipped.
- Table output lists foreign and state-owned brokers, and says when the
source rounded figures.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01V3syn2fN2LNo5pvTSqc6eh
Between the two latest releases there are 1,568 holder changes. With only
`limit` (up to 2000, ~200 KB), agents that wanted them all overflowed
their tool output: Hermes then paged ticker by ticker (32 calls), and
opencode 1 and 2 did the same and timed out.
- `offset` pages through the rows; responses carry `offset` and
`next_offset` (null after the last page). Pages cap at 300 rows
(~40 KB); the over-limit error points to `offset` and `summary`.
- Rows come largest |delta| first, so the first page holds the moves
that matter rather than tickers starting with A.
- `summary` covers every row: counts by change type, the number of
tickers, and the ten tickers with the most changes.
The CLI's `ownership changes` is unchanged. tools.json and the website's
MCP pages are regenerated for the new description and parameter.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01V3syn2fN2LNo5pvTSqc6eh
MSN's feed for a stock is often mostly general market news: of 10 items,
ADRO, BREN and CUAN had none about the company, BBRI one, yet they were
presented as the company's news. Fetch at least 50 items, label each
`relevance: company` (tagged by MSN with the stock's instrument id, or
naming the ticker or the company) or `market`, and return company items
first, newest first, trimmed to --limit.
Company names come from the quote's short and display names, best
effort: MSN tags Telkom articles with its NYSE ADR, not TLKM, so names
catch what tags miss. Classification stays a pure function in parse_news.
Output change, intentional: news JSON gains `relevance`, the table an
ABOUT column (goldens regenerated), and the get_company tool description
explains the field (tools.json and the website tools page regenerated).
News cache keys change so entries without the field are not reused.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01V3syn2fN2LNo5pvTSqc6eh
Findings from a Codex (gpt-6.1-sol, xhigh) review of v0.2.4..HEAD, each
verified in the code before fixing:
- high: ticker symbols reached cache file paths unchecked, so a symbol
like `../../REPORT` read (and deleted, if unparseable) .json files
outside the cache, reachable through MCP get_quotes. resolve_symbol now
accepts only ticker characters, and cache keys are encoded as single
safe path components.
- high: sync decided before downloading and replaced the database
without rechecking; an import that landed meanwhile was erased. It now
rechecks the local state before installing, and idx-mcp runs
ownership_sync exclusively with ownership reads.
- replaces_sqlite_sha256 never downgrades history: the client requires
the rebuild to keep at least the local release count, and the
publisher lists a replaced build only then.
- cache writes are atomic (temp file + rename), so concurrent readers
never see, and delete, a partial entry.
- idx-mcp caps symbols (20, deduplicated) and limits, and runs at most 4
tool calls at once.
- each Yahoo client gets its own cookie jar, paired with its crumb.
- ZIP entries (XLSX, KSEI archives) are read with a 256 MiB bound;
non-ASCII archive dates and huge percentages are parse errors instead
of panics or overflow.
- agent-smoke --keep deletes every copied agent config, not only auth
files.
CLI goldens unchanged; the MCP tool golden changes only to state the
new caps.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- tests/agents/agent-smoke.sh runs tests/agents/cases.json through
Claude Code, Codex, pi, Hermes Agent, and opencode headless against a
local idx-mcp build, each with an isolated config, and checks the
tools called, their key arguments, and (mock data) facts in the
answer. Opt-in; never part of the gate.
- crates/idx-mcp/tests/golden/tools.json pins tool names, descriptions,
and input schemas, so changes to what agents see are reviewed.
- nix develop .#agents adds Claude Code, Codex, and opencode from
nixpkgs after the host PATH.
First run: 30/30 (5 agents x 6 cases, mock data).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
