Missão 5 / Bloco Z.2: RELEASE_MANIFEST.json on this branch now describes
its own local migration state (head 0033_cfop_ippt_write_mode) instead of
mirroring master's deployed state (0032) - the mismatch was making
test_manifest_records_the_real_migration_head fail for a real but
uninteresting reason (this branch adds a migration file master doesn't
have). The note field is explicit that this is branch-local metadata, never
applied to any real Postgres, and that master's own manifest (not this one)
is the source of truth for what production actually runs. No merge to
master, no Agent change.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Missão 4 / Bloco E.1-E.2, on feature/cfop-ippt-write only (never main):
- Migration 0033_cfop_ippt_write_mode: Installation.cfop_ippt_write_mode
(DISABLED|DRY_RUN|APPROVAL_REQUIRED|ENABLED, DEFAULT 'DISABLED' for every
installation, existing and new) + append-only cfop_ippt_write_audits table.
- central/app/cfop_ippt_audit.py mirrors rtc_audit.py for this narrower
capability: job_authorizations(installation) (delivered inside the
authenticated job/heartbeat authorizations, never a local Agent switch) and
record_cfop_ippt_write_audits (persists applied/aborted cells; only CFOP/
IPPT are ever accepted, everything else silently ignored).
- central/app/web.py: POST /agents/{id}/cfop-ippt-write - ADMIN only (SSO
Gerenciador), CSRF, typing GRAVAR required for any mode other than
DISABLED, every change audited (CFOP_IPPT_WRITE_MODE). Small toggle UI in
agent_detail.html next to Contexto Fiscal.
- Wired into the existing job/heartbeat authorization payload and job
completion handling in central/app/api.py, alongside (never mixed with)
the RTC wiring.
- 18 new tests (central/tests/test_cfop_ippt_write_plumbing.py): default
DISABLED everywhere (job payload AND /v1/agents/me/authorizations); a
forged completion payload can never change the mode (read fresh from the
Installation row, never from Agent-supplied JSON); PADRAO role gets 403
with no audit row; ADMIN + wrong/missing CSRF is refused; ADMIN without
typing GRAVAR is refused (422); disabling never requires GRAVAR; invalid
mode is refused; the audit recorder only accepts CFOP/IPPT, silently
drops any other field (e.g. a stray CCLASSTRIB row).
- Two existing RTC tests updated for the additive authorizations-dict key
(cfop_ippt_write_mode) they now also carry.
Still NOT wired into agent/scripts/run_agent_service.py::execute_job - no
Agent code calls select_candidates/apply_candidates yet, so nothing writes
to Firebird from this branch. That wiring, an Agent version bump, a new
installer and Windows Sandbox validation are the next step, deliberately
left for after review (see docs/CFOP_IPPT_WRITE_CAPABILITY_V1.md) - this
mission explicitly asked not to publish this as an active release yet.
FULL_SUITE on this branch: 1 known, expected failure -
test_manifest_records_the_real_migration_head - because this branch adds
migration 0033 that was never merged to master/deployed; RELEASE_MANIFEST.json
intentionally still describes master/production (head 0032). Every other
test passes, including all CFOP/IPPT and RTC-adjacent tests.
REAL_PRODUCT_FISCAL_WRITES=0 (no write path is reachable from any Agent
code on any branch).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Adds test_weak_composed_evidence_never_becomes_a_candidate_regardless_of_the_write_flag,
proving from the real field_contract.resolve_product (not an isolated stub)
that the beer misclassification fixed on master (see
"fix: product_role never infers OWN_PRODUCTION from a single COMPOSICAO
occurrence") can never reach select_candidates as an AUTO_FILL_SAFE
candidate, independent of cfop_ippt_write_mode/write_gate.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Missão 4 / Bloco A.1 (prioridade máxima): central.app.field_contract.infer_role
treated ANY C000032.TIPO=COMPOSICAO row for a product as proof of own
production, with no ratio check - a beer sold 19 times and used once as a
recipe component (Agent 2, real data) was classified OWN_PRODUCTION from
that single occurrence, proposing to change its CFOP and IPPT.
New deterministic threshold in field_contract.py: COMPOSICAO must be a
majority (>=70%) of that product's own C000032 movement history, with at
least 5 total events, to count as own-production evidence. Below 30% the
evidence is discarded entirely (falls through to other role evidence,
e.g. TIPO=VENDA -> RESALE). Between 30% and 70% (with enough events to be
meaningful) is genuinely conflicting evidence -> UNKNOWN, never guessed
either way. RECEITA_INGREDIENTES/PROD_COZINHA-based evidence (a direct
recipe-table linkage, not a statistical ratio) is untouched.
agent/src/gti_fiscal_agent/firebird_reader.py now counts COMPOSICAO vs other
C000032 rows per product (read_product_evidence: single indexed query;
read_evidence: one grouped query for the whole catalog) into a new
ev["composed_ratio"] key, alongside the existing "composed" set (kept for
other, non-role consumers like coverage.py). central/app/enrichment/api.py's
sanitize_product_evidence and fiscal_completion.py's empty-ev fallback both
carry the new key through.
New tests (central/tests/test_field_contract.py) cover the 5 real cases named
in the mission: the beer (1 vs 19, 5% - discarded), a second beer (3 vs 4,
42.9% - MIXED/UNKNOWN), the liqueur (34 vs 22, 60.7% - still MIXED/UNKNOWN,
below the 70% bar), the juice with PROD_COZINHA=1+RECEITA_INGREDIENTES
(unaffected, stays OWN_PRODUCTION), and the tilapia/corvina fish (no
composed evidence at all, stays RESALE) - plus a strong 90% majority case
and an insufficient-sample (<5 events) case. A new end-to-end test in
agent/tests/test_cfop_ippt_auto_apply.py (committed separately, see the
feature/cfop-ippt-write branch) proves the weak-evidence beer never reaches
select_candidates as a write candidate, independent of any write flag.
Re-running the dry-run over Agent 2's real database (GTI.FDB, 527 products)
after this fix: AUTO_FILL_SAFE CFOP/IPPT cells drop from 12 to 3 (the two
fish CFOP fills + the juice's IPPT, all with strong, non-statistical
evidence) - docs/runs/GTI_dryrun_universe_summary_after_A1.json. This
corrects the Missão 3 report's 12/9(downgraded)/4(sustained) manual
judgment call with a deterministic, tested engine rule.
No schema/migration change. Zero Firebird writes (read-only analysis only).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Hub-only version bump for the static dossie_contador.md anonymization fix
and the Bloco B PIS/COFINS-source evidence update committed just before.
No schema/migration change, no Agent file touched.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Missao 3 / Bloco C.1-C.2: new agent/src/gti_fiscal_agent/cfop_ippt_auto_apply.py
mirrors the existing rtc_auto_apply.py safety model for a new, narrower
capability (CFOP_IPPT_WRITE_V1): compare-before-write, stale-write guard
("human override sempre vence" - a value changed since the candidate was
selected is never overwritten), one transaction per cell, readback
verification, and rollback that only restores a value still equal to what
was written. Candidates come exclusively from field_contract.resolve_product's
own AUTO_FILL_SAFE classification for CFOP/IPPT - no new fiscal rule.
14 new tests (agent/tests/test_cfop_ippt_auto_apply.py) prove: flag OFF means
the caller never reaches a write; only CHANGE+write_eligible CFOP/IPPT cells
become candidates; a human edit between decision and write blocks the write;
a second run over an already-applied cell writes zero; a technical failure on
one cell never stops the batch; rollback refuses when the value changed again.
Deliberately NOT wired into agent/scripts/run_agent_service.py::execute_job,
and the Hub-side authorization/audit plumbing (Installation.cfop_ippt_write_mode,
audit table, job-delivered authorization - the same pattern as
central/app/rtc_audit.py) does not exist yet. No currently-published Agent
behavior (1.3.8) changes with this commit - nothing imports this new module.
See docs/CFOP_IPPT_WRITE_CAPABILITY_V1.md for the acceptance roteiro (test
products only, never real ones) and what remains before this can ever be
activated for a real installation - deliberately left for review before any
further step, per this mission's explicit instruction not to publish it active.
Also: dossie_contador.md (internal + anonymized) updated with the Bloco B
finding that the ~9.5% CST=01 anomaly is confined to a single calendar month
(not the whole year), with no fiscal-dimension (NCM/CFOP/CSOSN/ICMS group)
explaining which items got it - consistent with a one-off technical event,
not a deliberate rule.
REAL_PRODUCT_FISCAL_WRITES=0 (no write path exists in this commit).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
New script (scripts/blocoB_pis_cofins_source.py) and evidence artifact
(docs/runs/BLOCOB_PIS_COFINS_SOURCE_V1.json): tests whether NCM, CODGRUPO, CFOP,
CSOSN or GRUPO_ICMS determine the CST_PIS_COFINS value emitted in NFCE_ITENS,
across all 5 installations. Read-only; no Firebird write path exists in this script.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Missao 2 (fechamento de lacunas fiscais, read-only + UI/engine gated):
- field_contract: pis_cofins_profile now also accepts the literal sentinels
NOT_APPLICABLE/OPERATION_TIME, resolving to NA only from an audited
FiscalContextEntry - never inferred from emission statistics. CFOP_ENTRADA
under an already-decided entry_scope=MIXED_OR_DYNAMIC now resolves to NA
instead of MANUAL, so it stops blocking product completeness and stops
re-asking an already-answered question. Regression tests prove: no ADMIN
decision => unchanged old behavior; decision present => gated bucket, no write.
- context_keys: validate_entry accepts the two sentinels for pis_cofins_profile.
- Contexto Fiscal UI: PIS/COFINS and CFOP_ENTRADA cards show the Bloco 2
cross-database evidence (5 installations, CRT 1 and 3: emission never reads
the PIS/COFINS cadastral fields anywhere) and a link to an ADMIN-facing,
anonymized dossie-for-the-accountant; entry_scope is now a dropdown.
- release/tests: decoupled the Hub's GTI_FISCAL_VERSION from the Agent's own
version (test_versions_are_coherent_across_hub_and_packages vs the new
test_agent_version_is_internally_coherent) - a Hub-only release must not
require touching, rebuilding or republishing the Agent.
- New read-only evidence: docs/runs/GTI_dryrun_universe_summary.json (527
products, Agent 2's real database), docs/runs/BLOCO2_PIS_COFINS_CROSS_DB_V1.json
(5 installations), docs/fiscal_completeness_matrix.md, docs/dossie_contador.md
(internal, real names) + central/static/docs/dossie_contador.md (public,
anonymized - no CNPJ/razao social/product names), docs/AGENT2_UPGRADE_RUNBOOK_V1.md.
- CLAUDE.md: recorded the permanent fiscal-mission rules (no AI/scraping, no
Firebird schema changes, deploy only via deploy_hub.ps1, never reenroll,
no remote access to client Agent hosts, no client data in unauthenticated
static routes).
No schema/migration change. Zero Firebird writes (new scripts never open a
write connection). Hub only - Agent code untouched.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Adds an ADMIN-only "Contexto Fiscal" section to the Agent detail page: confirm/alter/clear the fiscal decisions
field_contract.resolve_product genuinely cannot derive from the ERP alone (icms_treatment, pis_cofins_profile,
entry_scope). Persists through the existing FiscalContextEntry table (the same one blocker_resolution already
writes operator answers into) - no new persistence model, no opaque JSON box. CSRF + RBAC + audit reuse the exact
mechanism already proven for /agents/{id}/features. Confirming or clearing flags every already-enriched product on
the installation for re-evaluation.
The list was reduced from 5 candidates to 3 by auditing real production data (Agent 2, 2026-09-27/28):
icms_rate/credit_rate and rounding were dropped - both proven NOT_APPLICABLE / ALREADY_DERIVED for this CRT=1
company from real field_contract dry-runs, not assumption. Real evidence surfaced in the UI is a dated, one-time
manual read-only measurement (disclosed in the card copy itself as such) - the Hub has no direct Firebird access,
so this is not a live query; a follow-up should pipe it through the heartbeat the same way CRT/CONTRIBUINTE_IPI
were wired.
Also found, during the same audit, that this client's NFC-e emission ignores the product's cadastral PIS/COFINS
CST/rate fields entirely (verified on the 3 real products that DO have them populated - the emitted documents never
carry the cadastral value) - documented as a strong OPEN recommendation for a future field_contract change, not
applied here, since changing canonical fiscal semantics needs accountant sign-off, not just my own inference.
1129 tests passed (tests + central/tests + agent/tests), 0 failed. Version 1.3.9 (single GTI_FISCAL_VERSION). Hub
only - Agent unchanged, no new installer this release.
REPROCESS_PRODUCTS is not in SYNC_JOB_KINDS, so it fell to the else branch of execute_job and hardcoded company=None
for _unified_fiscal - a real gap, not a missing feature: the same deterministic, fail-closed company_context.resolve()
every sync job already uses is cheap (no full-table scan, confirmed against the real database) and applies just as
well to a one-off reprocess job. Now resolves the real company the same way; falls back to None only when the
company genuinely cannot be resolved yet, never a silent guess. Does not touch RTC rule logic.
Verified end-to-end against the real production database (dry-run, zero writes) that this was the last domain still
carrying the old bug - field_contract's post-NCM fiscal completion already bypassed it entirely by reading
Installation.company_uf/crt/company_contribuinte_ipi directly (populated via the 1.3.7 heartbeat fix).
1123 tests passed (tests + central/tests + agent/tests), 0 failed. Version 1.3.8 (single GTI_FISCAL_VERSION).
field_contract.resolve_product was only exercised by its own unit tests. /v1/products/resolve now calls it for every product whose NCM
resolves, merging CFOP/CST/CSOSN/PIS-COFINS/CST_IPI/SITUACAO_TRIBUTARIA/IAT/IPPT decisions alongside CLASSIFICACAO_FISCAL/CEST/IPI - reusing
the same engine, the same FiscalContextEntry table the blocker-resolution workflow already writes to, no duplicated rules.
Closes three real transport/discovery gaps found while wiring it in:
- Agent heartbeat never sent company_crt (env-var UF only) or company_contribuinte_ipi, so Installation.company_crt stayed permanently empty
even though company_context.resolve() already proves both from C000004/EMISSOR for every coverage job. Resolved once, cached, retried only
while unresolved. Migration 0032 adds Installation.company_contribuinte_ipi (NULL=unknown, never coerced to "N").
- The Agent's enrichment payload only ever reported 3 old fields (CLASSIFICACAO_FISCAL/CEST/IPI) as "current", so the Hub could never see the
product's own CFOP/CST/CSOSN/... nor the classification columns (TIPO/PROD_COZINHA/PRODUTO_SERVICO/CODGRUPO) infer_role needs - separated
into a read-only CURRENT_REPORT_FIELDS list, the write allowlist (PRODUCT_ENRICHMENT_WRITE_V1) untouched.
- Added a targeted, product_code-scoped evidence query (firebird_reader.read_product_evidence: recipes/composed/purchases/stock, indexed WHERE,
never a full-table scan) so st_withheld/product_role can be proven for a single ad-hoc resolve, not only during a full sync job.
Verified read-only against the real database (D:\PROJETOS\GTISISTEMAS\Banco\GTI.FDB): CRT=1, CONTRIBUINTE_IPI=N, company_context.resolve()
reaches HIGH confidence; infer_role correctly resolved RESALE/OWN_PRODUCTION for 5/8 sample real products with zero operator questions.
1121 tests passed (tests + central/tests + agent/tests), 0 failed. Version 1.3.7 (single GTI_FISCAL_VERSION shared by Hub and Agent per
central/app/version.py's own design - never independently versioned).
Sandbox test: a 26 s outage exhausted the previous 15 s retry budget and failed a REPROCESS_ALL.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Found by uninstall + fresh install on the clean Windows Sandbox: the old agent_state.sqlite identity conflicted with the new enrollment and the service sat idle.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Found in the clean-Windows recovery test: a single TLS handshake failure mid-job failed a whole sync.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
- resolver (rules RTC-R0..R4/A1..A3/G1..G2) over the official cClassTrib table 2026-06-22 and LC 214/2025 annexes; local vs official catalog comparison
- ONE root-cause RTC blocker per product (missing/unresolved/ambiguous/conflict/official mismatch/catalog outdated/evidence missing)
- auto-apply: C000025.CCLASSTRIB only, three switches (FIREBIRD_WRITES + RTC_AUTO_APPLY_ENABLED + Hub authorization), per-product re-read and re-decision, never overwrites, rowcount 1, audit table rtc_write_audits (migration 0027)
- copy-only write E2E (280/280 applied, only CCLASSTRIB changed, 2nd run 0 writes, stale decision 0 writes); real databases untouched
- docs: cclasstrib_resolution, rtc_auto_apply, unified_fiscal_completeness
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
- RTC: Agent reads ALL PARAMETROS rows every sync (conflict => off), reports rtc_source/rtc_observed_at/rtc_param_rows; legal_effective_context is informational only
- PIS/COFINS: applicability from CRT + PISCOFINS flag + incidence only (never from empty fields); Simples w/o special treatment NOT_APPLICABLE, taxed regime with explicit N OPTIONAL, unreadable declaration NEEDS_REVIEW
- CST: one digit with an official zero-padded form is VALID_AFTER_NORMALIZATION (non-blocking warning), nothing is written to the ERP
- tests: RTC toggle 0->1->0 across syncs (Agent + Hub stale closing), non-circularity, normalization, panel provenance
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
- central/app/unified_fiscal_coverage.py: pure rules (PIS/COFINS NOT_APPLICABLE/REQUIRED/OPTIONAL/NEEDS_REVIEW, RTC only when the ERP switch is on)
- agent: read-only loader (unified_fiscal.py) replaces the legacy/expanded module; findings feed the single operational state, coverage and blocker queue
- hub: BLOCKED_FISCAL_DATA blockers (PIS/COFINS/RTC) share the queue; stale ones close on the next full pass; dashboard breakdown gains CEST/PIS/COFINS/RTC
- panel: legacy/expanded coverage block removed; discreet 'Reforma Tributaria: Ativa/Desativada'
- docs: unified_fiscal_completeness, pis_cofins_applicability, rtc_pis_cofins_model + official source manifests; read-only E2E script
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
gti
