The ResetProgressFunc callback streamed per-phase messages from inside
the reset transaction, so a rollback could leave already-printed
progress describing work that never committed. Remove the callback
entirely; the final result summary printed by proboctl still conveys
the outcome without the misleading intermediate output.
Signed-off-by: Émile Ré <emile@probo.com>
When IP geolocation does not resolve a country, or resolves one with no
known cookie-consent regulation (common on localhost and unmapped
regions), the banner previously fell back to OPT_OUT with no recorded
regulation. Apply GDPR (OPT_IN) as the safe default in that case so the
strictest consent model wins when origin is unknown.
To keep consent records auditable, stamp each one with a regulation
source of DETECTED (resolved from geolocation) or DEFAULT (fell back to
GDPR). The shared cookiebanner.ResolveRegulation helper centralizes the
decision for both the config and consent endpoints, and the new value is
exposed through GraphQL, MCP, the CLI, the n8n node, and the console
consent-records views.
Signed-off-by: Émile Ré <emile@probo.com>
The preference panel is a fixed, backdrop-less card whose only
scrollable region is the category list. Wheel events over the
header, footer, padding, or at the list edges chained to the host
page, so the page scrolled instead of the panel.
Treat the open panel as the modal it already claims to be: lock
document scroll while state is "panel" (compensating for scrollbar
width to avoid layout shift) and add overscroll-behavior: contain
to the list. Restore scroll on close and on disconnect so the page
is never left frozen.
Signed-off-by: Émile Ré <emile@probo.com>
A pre-fix bug in the cookie-banner SDK walked its own bundle frame
when computing a tracker's initiator, so cookies and storage written
by third parties, malware, or browser extensions were attributed to
cookie-banner.iife.js. The report upsert keeps the existing value on
re-detection (initiator_url = COALESCE(new, old)), so rows whose
corrected initiator is NULL would keep the stale bundle URL forever.
Null initiator_url and initiator_domain on detected_trackers rows that
point at the @probo/cookie-banner bundle. Genuine third-party rows
repopulate the correct initiator on the next detection.
Signed-off-by: Émile Ré <emile@probo.com>
The cookie/storage/resource detectors only skipped the API base-url
origin and the page origin when computing a tracker's initiator. When
the SDK is served from a CDN (e.g. jsDelivr) distinct from the API
host, its own wrapper frames sit atop every setItem/document.cookie
call stack and were never skipped, so getInitiatorURL returned the
bundle URL. Third-party and browser-extension writes were therefore
misattributed to cookie-banner.iife.js.
Capture the SDK's own served script URL once at load (currentScript
src, with an Error().stack fallback) and exclude it, at URL level, from
both the initiator stack-walk and the resource detector. URL-level
exclusion keeps other trackers served from the same shared CDN
detectable.
Signed-off-by: Émile Ré <emile@probo.com>
Operators previously had no way to curate the global tracker catalog
beyond inspection and banner-scoped resets. Add three proboctl commands
backed by small coredata helpers:
- common-third-party upsert: create or update a vendor keyed by slug,
with partial-merge so an unset flag never blanks an existing column.
- common-tracker-pattern link/unlink: repoint catalog rows at a common
third party (re-arming enrichment and remapping the uncategorised org
trackers so the mapping worker re-resolves the vendor) or detach them.
Unlinking skips enrichment and remap since there is no new vendor.
- common-tracker-pattern set-description: write a description, mark the
row enriched, and backfill linked org patterns lacking one.
Signed-off-by: Émile Ré <emile@probo.com>
The cookie-database aggregator backstop normalised the agent's
third-party name and looked it up against bare brand keys, but
normalizeAlnum folds the eTLD into the key (cookiedatabase.org ->
cookiedatabaseorg). Domain- and URL-form attributions therefore
slipped past the exact lookup, letting noisy aggregator names be
accepted instead of discarded.
Add uri.DomainLabel to reduce a host-like string to its primary
registrable label and check it alongside the normalised name, so
both brand ("Cookiepedia") and domain forms ("cookiedatabase.org",
"https://www.cookiepedia.co.uk/list") resolve to the same key.
Signed-off-by: Émile Ré <emile@probo.com>
Cookie-database and consent-directory sites (Cookifi, Cookiepedia,
cookiedatabase.org, CookieServe, ...) rank highly in web search only
because they catalog cookies, not because they set them. The mapping
agent could attribute a tracker to the directory operator itself
instead of the vendor the page names.
Instruct the agent to treat such results as reference directories and
extract the named vendor, never the operator, while keeping a CMP's own
product cookie attributable (OptanonConsent -> OneTrust, CookieConsent
-> Cookiebot). Add a conservative code backstop that discards
attributions to pure aggregators, scoped to exclude CMP vendors so
legitimate own-cookie attributions survive.
Signed-off-by: Émile Ré <emile@probo.com>
Patterns like "ethereum-https://letaido.com" embed the scanned site's
own origin, appended either by a browser extension (e.g. a wallet
injecting window.ethereum) or by an owner-set tracker. The mapping
agent read that embedded domain and attributed the site to itself as a
third party ("Letaido"), but a site owner is never a third party of its
own site.
Pass the scanned site's eTLD+1 to the agent as a <scanned_site> signal
and instruct it to ignore a domain embedded in the pattern that matches
the site, returning low confidence when that is the only cue. Add a
deterministic backstop that discards any agent attribution whose vendor
name resolves to the site's own domain, so the pattern falls through to
the unmatched fallback regardless of whether the model honored the
prompt. Deterministic catalog, naming, and sibling signals run before
the agent and are untouched, so an independent strong vendor signal
still attributes normally.
Signed-off-by: Émile Ré <emile@probo.com>
The reset-trackers operator command reset every uncategorised,
non-excluded pattern of a banner and printed only a single summary
line once the transaction committed, giving no feedback during long
rebuilds.
Add a --keyword flag that scopes both the glob decomposition and the
mapping reset to patterns whose pattern or display name contains the
substring. The match lives in a new TrackerPatternFilter.WithPatternKeyword
field so it runs in SQL and is shared by the glob load and the
ResetAndRequestMappingByCookieCategoryID update, keeping the two in
lockstep. The banner-wide pattern-analysis re-arm is left unscoped.
Thread an optional progress callback through ResetBannerTrackers so the
command streams per-phase updates (category load, per-glob decomposition,
mapping reset, analysis re-arm) as the work runs.
Signed-off-by: Émile Ré <emile@probo.com>
The reenrich command forced callers to pass exactly one selection
anchor (--id, --linked-banner, --linked-org, or --common-third-party)
before any pattern could be re-described. That made the common case of
re-enriching every catalog row with a blank description impossible
without an artificial anchor.
Drop the anchor-required guard while keeping anchors mutually
exclusive. With no anchor, the filtering flags now select across the
whole catalog, so --without-description re-enriches every pattern that
lacks a description.
Signed-off-by: Émile Ré <emile@probo.com>
The "Detected Count" label was ambiguous about whether it counted
distinct trackers or report occurrences. The report endpoint upserts
each detected tracker by identifier, so the value is the number of
distinct trackers matched by the pattern. Rename the label to
"Distinct Trackers Detected" to reflect this.
Signed-off-by: Émile Ré <emile@probo.com>
Replace the many-to-many junction table with a direct
parent_third_party_id foreign key on third_parties. Each
sub-third-party now belongs to exactly one parent, making
duplicates across parents independent entities.
Replace the firstLevel boolean with an integer level field
(1 = direct, 2+ = parent level + 1) to support arbitrary
nesting depth.
Remove the createThirdPartyThirdPartyMapping and
deleteThirdPartyThirdPartyMapping mutations, the CLI
link/unlink commands, and the corresponding MCP tools.
Creating a child third party now just requires passing
parentThirdPartyId on the existing createThirdParty mutation.
The frontend walks the parentThirdParty chain to build
display names like "Name (Ancestor1/Ancestor2)" and shows
clickable ancestor links on the detail page.
Signed-off-by: Sacha Al Himdani <sacha@getprobo.com>
Register Neon as a connector provider and add a new access review
driver that fetches organization members from the Neon API with
cursor-based pagination.
Neon's OAuth is partner-gated, so the connector is API-key only
(Bearer, the default scheme). A personal or organization API key can
belong to several organizations; the operator supplies the ID of the
one to review. The members endpoint exposes per-user MFA state
(has_mfa) and deactivation, which map to the access entry MFA status
and active flag; the stable account UUID (user_id) is used as the
external ID over the membership ID.
Signed-off-by: Aurélien Sibiril <81782+aureliensibiril@users.noreply.github.com>
Register Render as an API-key connector provider and add an access
review driver that fetches workspace members from the Render API
(GET /v1/owners/{ownerId}/members).
Render exposes no partner OAuth program, so the connector authenticates
with a read-scoped API key (Authorization: Bearer) plus the customer's
Workspace ID. The flat members endpoint reports an explicit account
status and MFA flag, surfaced as the Active and MFAStatus fields; the
stable "usr-" id becomes ExternalID. There is no picker -- the
workspace is captured up front via ExtraSettings -- so
SetOrganizationSettings is omitted.
Signed-off-by: Aurélien Sibiril <81782+aureliensibiril@users.noreply.github.com>
Register Qovery as a connector provider and add a new access review
driver that fetches organization members from the Qovery API.
Extend API key connection handling with a configurable Authorization
token scheme so Qovery can use "Token" while existing providers
continue to default to Bearer.
Signed-off-by: Cursor Agent <cursoragent@cursor.com>
Co-authored-by: Bryan FRIMIN <bryan@frimin.fr>
Replace the FileSign field in api.Config and server.Config with File
(*file.Service). Pass the new file.Service and other required deps to
files_v1.NewMux. Remove the now-redundant filesign package. Update
the favicon URL path to /api/files/v1/public/ in server.go.
Signed-off-by: Ludovic Vielle <ludovic@probo.com>
Add GET /public/{fileID} (unauthenticated, public files only) and
GET /{fileID} (session+API key+OAuth2, all files, core:file:get IAM
check). IAM and not-found errors both return 404 to prevent leaking
whether a file exists.
Signed-off-by: Ludovic Vielle <ludovic@probo.com>
The two actions expressed the same permission. Consolidate on
core:file:get and remove the now-redundant core:file:download-url
constant, policy entries, and all three call sites.
Signed-off-by: Ludovic Vielle <ludovic@probo.com>
Replace the inline PresignGetObject logic in probo.FileService with a
call to fileManager.GenerateFileUrl. Update the two callers.
Signed-off-by: Ludovic Vielle <ludovic@probo.com>
Add GeneratePublicPresignedURL and GeneratePresignedURL methods to
pkg/file. Update the GenerateFileURL path to /api/files/v1/public/
to reflect the upcoming endpoint rename. Pass fileManagerService to
file.NewService in probod.
Signed-off-by: Ludovic Vielle <ludovic@probo.com>
Move connectorProviderInfos from Organization to a new root query field
accessReviewDrivers, backed by a deployment-scoped policy so any
authenticated identity can list it without an org-scoped permission check.
Delete the now-unused helper file and update the frontend and e2e tests.
Signed-off-by: Bryan Frimin <bryan@probo.com>
Better Stack exposes team members and pending invitations through its
Uptime API. Wire it as an access-review connector so a Better Stack
team can be reviewed in access-review campaigns.
Better Stack has no third-party OAuth app for listing members (its
OAuth is an end-user MCP sign-in), so the connector authenticates with
a Bearer API token plus the team name that scopes the team-members
listing. The driver paginates /api/v2/team-members, maps roles and
invitation records into account records, and the source name is
resolved from the configured team.
This wires the full surface: the provider enum and migration, the
connector settings, the registry registration with the team-name extra
setting, the GraphQL input and resolver marshaling, the frontend field
mapping and connector logo, and cassette-backed driver tests.
Signed-off-by: Aurélien Sibiril <81782+aureliensibiril@users.noreply.github.com>
NewCursorFromFlags validated --first/--last, --after/--before, and
--after/--last as mutually exclusive, but missing the symmetric
--first/--before guard. Passing --first with --before fell into the
backward branch, where size defaulted from the unset --last and the
direction silently flipped to Tail, discarding --first and reversing
pagination. Reject the combination up front like the other invalid
pairings.
Signed-off-by: Émile Ré <emile@probo.com>
RegenerateCookieBannerTrackerPolicy authorized against the generic
cookie-banner update action, conflating policy regeneration with banner
edits. Add a dedicated regenerate-policy action and authorize both the
console and MCP resolvers against it so the capability can be granted
independently.
Signed-off-by: Émile Ré <emile@probo.com>
Replace the limit-driven auto-walking Paginate helper with explicit
cursor-pagination flags (--first/--after, --last/--before) that mirror
the GraphQL connection arguments. List commands now return a single
keyset page with its page info, and emit cursors so callers can page
forward and backward. --before no longer requires --last: both --first
and --last default to 50 when omitted.
Also split the tracker-pattern stats into enriched with and without a
description so the enrichment backlog is visible at a glance.
Signed-off-by: Émile Ré <emile@probo.com>
The common tracker pattern show command swallowed every error from
loading the linked common third party, hiding real database failures
behind silent partial output. Only tolerate ErrResourceNotFound now and
wrap any other error so the command fails loudly.
In the console, the copy-to-clipboard button reported success
unconditionally. Await the clipboard write and surface an error toast
when it rejects, guarding against a missing common tracker id.
Signed-off-by: Émile Ré <emile@probo.com>
Drop the --org mode and the --banner flag: the command now takes the
banner GID as a positional argument and resets exactly one banner. The
org-wide fan-out added operational risk for a destructive action without
a clear use case, and a positional GID reads more naturally than a flag.
Signed-off-by: Émile Ré <emile@probo.com>
The --reset-enriched flag was effectively a no-op: the enrichment
worker claims rows solely on enrichment_requested_at, and SetEnriched
rewrites enriched_at regardless, so clearing it never changed whether a
row was re-processed. Remove the flag and the resetEnriched parameter on
RequestEnrichmentByIDs, which now only stamps enrichment_requested_at.
Add a --without-description filter to the list and reenrich commands,
backed by a new described predicate on CommonTrackerPatternFilter, so an
operator can target catalog rows that still lack a description.
Signed-off-by: Émile Ré <emile@probo.com>
Require exactly one selection anchor (--id, --linked-banner,
--linked-org, or --common-third-party) for common-tracker-pattern
reenrich, dropping the catch-all --all; the tracker-type, keyword, and
state flags now narrow the anchor's result except when explicit --id
values are given. Add --linked-banner, --linked-org, and a tri-state
--with-common-third-party to the list command, replacing the separate
--linked/--unlinked booleans, and rename --third-party to
--common-third-party across both commands.
Support these by adding an ID restriction to CommonTrackerPatternFilter
so linked-banner/linked-org selections can be intersected with the
remaining filters in a single query.
Memoize the pg client on the proboctl Factory to avoid a duplicate
Prometheus collector registration panic when more than one command
path builds a client. Surface timestamps in both listing tables and
flag enriched-but-undescribed rows in the displayed enrichment state.
Signed-off-by: Émile Ré <emile@probo.com>
The tracker-mapping, common-pattern enrichment, and third-party
disambiguation agents were all built from one shared
TrackerAgentsConfig fed by a single tracker-mapping config slot. That
forced a single AgentTimeout to be reused and patched per worker, and
two unrelated max-turns fields to share one struct.
Split the in-code config into TrackerMappingAgentConfig,
TrackerEnrichmentAgentConfig, and DisambiguationAgentConfig, each with
its own timeout and max-turns, and add dedicated tracker-enrichment and
third-party-disambiguation provider slots (the latter resolving next to
third-party-vetter). Enrichment and disambiguation fall back to the
tracker-mapping slot when their own provider is unset, preserving
single-config deployments.
Drop the shared pkg/agentsbuild package and duplicate its small wiring
into probod and proboctl so the two executables stay decoupled. Wire
the new env vars, builder test coverage, and Helm values.
Signed-off-by: Émile Ré <emile@probo.com>
