Releases: ArabAgentSkills/arab-payments-skill-atlas
Release list
Arab Payments Skill Atlas v1.4.2
Patch release following the v1.4.1 MyFatoorah refund source-watch update.\n\nChanges:\n- Updated the Python installer to use GH_TOKEN or GITHUB_TOKEN for latest-release GitHub API and zipball downloads.\n- Prevents authenticated maintainer runs from failing on unauthenticated GitHub API rate limits.\n- Added regression coverage for authenticated updater request headers.\n\nValidation passed before release: validate_skill, validate_adapters, validate_source_watch, secret_scan, check_source_links, check_source_changes --check, eval scenario validation, unit tests, installer dry-run, and git diff --check.
Arab Payments Skill Atlas v1.4.1
Patch release for reviewed MyFatoorah refund source-watch findings from private watcher issue #18 and run 34112647995.\n\nChanges:\n- Added current MyFatoorah refund source coverage for the Refunds guide, Create Refund, Get Refund Details, and Refund Status Changed webhook model.\n- Clarified that Create Refund creates a refund request, not final refunded settlement.\n- Added guidance to rely on Refund Status Changed webhooks, use Get Refund Details as fallback, compare refund amount/account base currency/PaymentId, and block duplicate partial-refund retries.\n- Refreshed public-safe source-watch baseline/report metadata and added an eval guard for refund status handling.\n\nValidation passed locally before release: validate_skill, validate_adapters, validate_source_watch, secret_scan, check_source_links, check_source_changes --check, installer dry-run, unit tests, and git diff --check.
v1.4.0
Changes
- Updated Tabby guidance for current UAE/AED and Saudi Arabia/SAR custom integration scope, nested hosted-checkout URL extraction, merchant-code webhook registration, and duplicate-capture race handling.
- Added Tamara pre-check eligibility guidance before offering Tamara or creating the final checkout session.
- Added HyperPay handling for result code
000.000.001so partial approvals do not trigger full-order fulfillment without approved-amount and policy checks. - Hardened source-watch normalization for provider docs Markdown URL hints and Mintlify command-palette chrome.
- Added eval guards for Tabby capture races, Tamara eligibility pre-checks, and HyperPay partial approvals.
Validation
python -m unittest discover -s testspython .\scripts\validate_skill.pypython .\scripts\validate_adapters.pypython .\scripts\validate_source_watch.pypython .\scripts\secret_scan.pypython .\scripts\check_source_links.pypython .\scripts\check_source_changes.py --check
Arab Payments Skill Atlas v1.3.0
Summary
- Clarifies Tamara auto-authorisation guidance: confirmed auto-authorisation reaches
authorised, not capture or settlement. - Keeps capture as a separate fulfillment/settlement step unless current merchant-specific docs or account settings prove otherwise.
- Adds an eval guard against treating Tamara auto-authorisation as immediate capture or settlement.
- Refreshes public-safe source-watch metadata after private watcher issue #16 review through run 32000027072.
- Hardens source-watch GitHub metadata checks to use
GITHUB_TOKENorGH_TOKENwhen available.
Validation
python .\scripts\validate_skill.pypython .\scripts\validate_adapters.pypython .\scripts\validate_source_watch.pypython .\scripts\secret_scan.pypython .\scripts\check_source_links.pypython .\scripts\check_source_changes.py --checkpython .\scripts\install_or_update_skill.py --agent codex --dry-runpython -m unittest discover -s testsgit diff --check
Arab Payments Skill Atlas v1.2.2
Public-safe source-watch maintenance release.\n\n- Updated Tamara monitored source mapping to the provider's current webhook registration and authorise-order guide/API pages from the official docs index.\n- Hardened source-watch normalization for provider docs feedback/footer chrome.\n- Refreshed public-safe source-watch metadata after maintainer review.\n- No payment behavior guidance changes; endpoint-level details still require current official provider docs.
v1.2.1 - Tap source watch metadata
Summary
- Replaced Tap Payments monitored source URLs with the official Markdown variants advertised by Tap's
llms.txtindex. - Refreshed public-safe source-watch baseline/report after private watcher issue #14 and run 29230960823.
- No payment behavior guidance changed; reviewed Tap, Geidea, MyFatoorah, Tabby, and Tamara diffs as source URL, docs-index, code-example, or formatting churn.
Validation
python .\scripts\validate_skill.pypython .\scripts\validate_adapters.pypython .\scripts\validate_source_watch.pypython .\scripts\secret_scan.pypython .\scripts\check_source_links.pypython .\scripts\check_source_changes.py --checkpython .\scripts\install_or_update_skill.py --agent codex --dry-rungit diff --checkpython -m unittest discover -s tests
v1.2.0
Public-safe source-watch release for reviewed BNPL documentation changes.
- Updated Tabby guidance for payment webhooks vs dispute webhooks and the current 180-day refund initiation window.
- Updated Tamara guidance for authorisation/capture flow and current 21-day auto-capture handling for authorised but uncaptured orders.
- Added focused eval coverage for Tabby dispute/refund handling and refreshed source-watch metadata.
Validation passed before release: validate_skill, validate_adapters, validate_source_watch, secret_scan, check_source_links, check_source_changes --check, install_or_update_skill --agent codex --dry-run, git diff --check, and source-watch unit tests.
v1.1.3
Patch source-watch maintenance release.\n\n- Hardened source-watch normalization for docs copy-action chrome, EasyKash-style relative-age text, and GitHub repository updated_at metadata churn.\n- Refreshed the public-safe source-watch baseline/report after private watcher review for issue #12.\n- No provider payment behavior guidance changed.
v1.1.2
Summary
- Hardened source-watch comparison so OK-to-JavaScript-challenge degradations are treated as maintainer review warnings, matching the source-link checker.
- Treated transient HTTP 408, 429, and 5xx fetch failures as review warnings while preserving non-transient HTTP failures as source changes requiring review.
- Added regression tests for runner-only GitHub API JavaScript challenges and transient HTTP fetch failures.
Classification
transient_fetch/ watcher hardening for GitHub runner-only Kashier GitHub API 403 JavaScript challenges.- No new endpoint, auth/signature, webhook, status, amount/currency, refund/capture/void, test-card, adapter, eval, or payment-behavior guidance change.
Validation
python .\scripts\validate_skill.pypython .\scripts\validate_adapters.pypython .\scripts\validate_source_watch.pypython .\scripts\secret_scan.pypython .\scripts\check_source_links.pypython .\scripts\check_source_changes.py --checkpython .\scripts\install_or_update_skill.py --agent codex --dry-runpython .\scripts\eval_scenarios.py --validate --require-automated-checkspython -m unittest discover -s testsgit diff --check
v1.1.1
Summary
- Hardened source-watch normalization to ignore provider docs navigation/sidebar/footer chrome, AI-readable index notices, GitBook Markdown notices, recent-request widgets, and approximate relative update ages before hashing.
- Added regression tests for those chrome-noise cases so future watcher runs do not trigger release review for non-payment documentation chrome churn.
- Refreshed the public-safe source-watch baseline/report after the improved normalizer.
Classification
chrome_noise/ watcher hardening.- No new public payment behavior guidance beyond the already released Moyasar
v1.1.0update.
Validation
python .\scripts\validate_skill.pypython .\scripts\validate_adapters.pypython .\scripts\validate_source_watch.pypython .\scripts\secret_scan.pypython .\scripts\check_source_links.pypython .\scripts\check_source_changes.py --checkpython .\scripts\install_or_update_skill.py --agent codex --dry-runpython .\scripts\eval_scenarios.py --validate --require-automated-checkspython -m unittest discover -s testsgit diff --check