bondery-legal

النجوم١٦

التفرعات٠

آخر تحديث١ أغسطس ٢٠٢٦ في ١٢:٤١

Bondery legal disclosure hygiene — subprocessor registry, policy-claims reconciliation, data-flow triggers, and escalation boundaries. Use when adding third-party vendors/SDKs, new data collection, analytics, AI tools, retention/deletion, or editing Privacy/Terms.

التثبيت

التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.

المصدر

usebondery

usebondery/bondery

فتح مستودع GitHub عرض مستودعات المنشئ

تنزيل

المهن ذات الصلةSOC

استنادا إلى تصنيف SOC المهني

محللو أنظمة الحاسوبمهن الحاسوب والرياضيات·SOC 15-1211

مستكشف الملفات

8 ملفات

SKILL.md

readonly

name	bondery-legal
description	Bondery legal disclosure hygiene — subprocessor registry, policy-claims reconciliation, data-flow triggers, and escalation boundaries. Use when adding third-party vendors/SDKs, new data collection, analytics, AI tools, retention/deletion, or editing Privacy/Terms.
metadata	{"version":"1.0.0","namespace":"bondery"}

Bondery Legal

This skill reconciles product behavior with published legal claims. It does not certify regulatory compliance, draft final contractual language, or replace legal counsel. When in doubt, flag and escalate to a human owner.

When to use

Adding a new third-party SDK, API client, or npm dependency that sends data off-platform
Adding or changing a data-collecting feature (new PII field, tracking event, AI tool call, webhook)
Touching account deletion, backup, retention, or log-anonymization code
Touching analytics opt-out, cookie behavior, or DO_NOT_TRACK
Editing apps/website/src/components/legal/Privacy.tsx or Terms.tsx
Reviewing a PR that touches any of the above

Do not activate for routine UI refactors, API contract work, or technical security enforcement — use bondery-api, bondery-ux, or bondery-security instead.

Reconciliation workflow

Classify the change — does it add/remove a vendor, add/remove a data category, or change retention/consent behavior? If none, skip this skill.
Check the subprocessor registry — references/subprocessor-registry.md. New vendor not listed → blocking flag.
Check the policy claims inventory — references/policy-claims-inventory.md. Does the change satisfy, contradict, or leave orphaned a claim in Privacy.tsx / Terms.tsx?
Classify the fix:
- Factual correction (wrong vendor name, feature now exists) → agent may propose copy diff, labeled reality-sync
- Substantive legal claim (retention window, rights language, liability, jurisdiction) → draft suggestion only; tag for human/counsel sign-off
Update the registry in the same PR as the vendor/data-flow change.
Complete the legal checklist before merge.

Full workflow with examples: references/data-flow-workflow.md.

Non-negotiables (ranked)

Never assert regulatory compliance — no "GDPR/CCPA/SOC2 compliant" in code, commits, PR descriptions, or docs
New off-platform vendor sending user or contact PII must be checked against the subprocessor registry before merge — undeclared vendor is a blocking flag
Never edit substantive legal wording unsupervised — retention commitments, rights language, liability, jurisdiction, age minimums require human/counsel review; factual corrections (wrong vendor name) may be proposed as reality-sync only
New data-collecting features must be checked against existing policy claims — do not ship silently when claims would be contradicted or left orphaned
Self-host ≠ cloud — never imply the hosted Privacy Policy's subprocessors or retention commitments apply to a self-hosted deployment (references/self-host-vs-cloud.md)
Retention/deletion claims must trace to code — flag numeric claims (30-day backup purge, 90-day IP logs) with no enforcing job as gaps, not assumed targets
AI scope must not silently expand — disclosed use is "only when you message the AI Assistant"; new LLM paths sending contact PII require policy update
Never guess vendor data practices — cite the vendor's own privacy policy or DPA URL
Escalate ambiguity — surface the question to a human owner; do not adjudicate

Escalation rules: references/escalation-boundaries.md.

Legal document locations

Document	Path	Status
Privacy Policy	`apps/website/src/components/legal/Privacy.tsx`	Published (last updated July 13, 2026)
Terms of Service	`apps/website/src/components/legal/Terms.tsx`	Draft placeholder — not final contractual terms
Layout helper	`apps/website/src/components/legal/shared/LegalDocumentLayout.tsx`	—
Routes	`apps/website/src/app/(marketing)/(legal)/{privacy,terms}/page.tsx`	—
App links	Webapp login, mobile settings/legal	External links to website URLs

No docs/legal/ folder. Legal copy is English-only React components on the website.

Decision tree

Task	Read
Vendor / subprocessor list (code-grounded)	references/subprocessor-registry.md
Policy claim vs. code status	references/policy-claims-inventory.md
Trigger → action workflow	references/data-flow-workflow.md
Self-host vs. cloud obligations	references/self-host-vs-cloud.md
What agent may draft vs. must escalate	references/escalation-boundaries.md
Future: manifest, CI, roadmap	references/long-term-roadmap.md

Full index: references/README.md.

Cross-skill owners: technical enforcement → bondery-security; API contracts → bondery-api; UI/copy display → bondery-ux.

Legal checklist (before merge)

Change classified: vendor / data category / retention / consent / policy copy
Subprocessor registry checked — new vendor added or policy-sync status updated in same PR
Policy claims inventory checked — no silent contradiction with Privacy.tsx / Terms.tsx
Fix classified: reality-sync (factual) vs. escalate (substantive legal wording)
No regulatory compliance assertions in code, commits, or PR text
Self-host vs. cloud context considered if deploy/docs touched
AI data flows stay within disclosed "message the AI Assistant" boundary
Vendor privacy/DPA URL cited — not inferred
Ambiguity flagged to human owner — not resolved by agent alone

المزيد من هذا المستودع

نفس المستودع

bondery-changelog

usebondery/bondery

Bondery product changelog format, change categories, commit prefixes, calendar versioning, and release-note workflow. Use when editing docs/changelog/unreleased.mdx, documenting a user-visible or operationally notable change, preparing a release, choosing a conventional commit prefix, recording dependency updates, or drafting release communications.

2026-08-0216

bondery-release

usebondery/bondery

Bondery release operator runbook — prerequisites on main, extension gates, promote main:release and product tags, Dokploy image pins, rollback, and hotfix sequencing. Use when cutting a monthly or patch release, pushing to release, tagging api-X.Y.Z / webapp-X.Y.Z / ext-X.Y.Z, pinning image tags, promoting to production, or rolling back.

2026-08-0216

bondery-verification-loop

usebondery/bondery

Risk-proportional verification for Bondery monorepo changes. Maps changed files to npm workspaces and shared-package consumers, runs the fastest relevant lint, type, policy, test, contract, build, and runtime checks, then reviews staged and unstaged diffs without requiring commits. Use after implementing or refactoring, before handoff or a PR, when asked to verify/test/check changes, or when deciding which Bondery quality gates apply.

2026-08-0116

bondery-security

usebondery/bondery

Bondery security patterns — tenant isolation, auth areas, sessions, secrets, input validation, uploads, webhooks, client trust boundaries, and deployment hardening. Use when adding authentication, authorization, API routes, secrets, uploads, webhooks, payments, sensitive data, extension bridges, or reviewing security-sensitive changes.

2026-08-0116

bondery-api

usebondery/bondery

Bondery Fastify API contracts — transport wrappers, list pagination (offset/search/sort), resource-keyed responses, mutations, Stripe-style errors, route registration order, mobile sync, rate limits, and versioning. Use when adding or changing API routes, client API calls, OpenAPI, sync protocol, or API error codes.

2026-08-0116

bondery-skill-authoring

usebondery/bondery

How to create, name, and structure Bondery agent skills per agentskills.io. Use when adding or updating skills in .agents/skills/, writing SKILL.md frontmatter, progressive disclosure, references/, scripts/, or pre-ship checklists for Bondery skills.

2026-07-3116

Bondery Legal

This skill reconciles product behavior with published legal claims. It does not certify regulatory compliance, draft final contractual language, or replace legal counsel. When in doubt, flag and escalate to a human owner.

When to use

Adding a new third-party SDK, API client, or npm dependency that sends data off-platform

Adding or changing a data-collecting feature (new PII field, tracking event, AI tool call, webhook)

Touching account deletion, backup, retention, or log-anonymization code

Touching analytics opt-out, cookie behavior, or DO_NOT_TRACK

Editing apps/website/src/components/legal/Privacy.tsx or Terms.tsx

Reviewing a PR that touches any of the above

Do not activate for routine UI refactors, API contract work, or technical security enforcement — use bondery-api, bondery-ux, or bondery-security instead.

Reconciliation workflow

Classify the change — does it add/remove a vendor, add/remove a data category, or change retention/consent behavior? If none, skip this skill.

Check the subprocessor registry — references/subprocessor-registry.md. New vendor not listed → blocking flag.

Check the policy claims inventory — references/policy-claims-inventory.md. Does the change satisfy, contradict, or leave orphaned a claim in Privacy.tsx / Terms.tsx?

Classify the fix:

Factual correction (wrong vendor name, feature now exists) → agent may propose copy diff, labeled reality-sync
Substantive legal claim (retention window, rights language, liability, jurisdiction) → draft suggestion only; tag for human/counsel sign-off

Update the registry in the same PR as the vendor/data-flow change.

Complete the legal checklist before merge.

Non-negotiables (ranked)

Never assert regulatory compliance — no "GDPR/CCPA/SOC2 compliant" in code, commits, PR descriptions, or docs

New off-platform vendor sending user or contact PII must be checked against the subprocessor registry before merge — undeclared vendor is a blocking flag

Never edit substantive legal wording unsupervised — retention commitments, rights language, liability, jurisdiction, age minimums require human/counsel review; factual corrections (wrong vendor name) may be proposed as reality-sync only

New data-collecting features must be checked against existing policy claims — do not ship silently when claims would be contradicted or left orphaned

Self-host ≠ cloud — never imply the hosted Privacy Policy's subprocessors or retention commitments apply to a self-hosted deployment (references/self-host-vs-cloud.md)

Retention/deletion claims must trace to code — flag numeric claims (30-day backup purge, 90-day IP logs) with no enforcing job as gaps, not assumed targets

AI scope must not silently expand — disclosed use is "only when you message the AI Assistant"; new LLM paths sending contact PII require policy update

Never guess vendor data practices — cite the vendor's own privacy policy or DPA URL

Escalate ambiguity — surface the question to a human owner; do not adjudicate

Legal document locations

Document

Path

Status

apps/website/src/components/legal/Privacy.tsx

Published (last updated July 13, 2026)

apps/website/src/components/legal/Terms.tsx

Draft placeholder — not final contractual terms

Layout helper

apps/website/src/components/legal/shared/LegalDocumentLayout.tsx

—

Routes

apps/website/src/app/(marketing)/(legal)/{privacy,terms}/page.tsx

—

App links

Webapp login, mobile settings/legal

External links to website URLs

No docs/legal/ folder. Legal copy is English-only React components on the website.

Decision tree

Task

Read

Vendor / subprocessor list (code-grounded)

references/subprocessor-registry.md

Policy claim vs. code status

references/policy-claims-inventory.md

Trigger → action workflow

references/data-flow-workflow.md

Self-host vs. cloud obligations

references/self-host-vs-cloud.md

What agent may draft vs. must escalate

references/escalation-boundaries.md

Future: manifest, CI, roadmap

references/long-term-roadmap.md

Cross-skill owners: technical enforcement → bondery-security; API contracts → bondery-api; UI/copy display → bondery-ux.

Legal checklist (before merge)

Change classified: vendor / data category / retention / consent / policy copy

Subprocessor registry checked — new vendor added or policy-sync status updated in same PR

Policy claims inventory checked — no silent contradiction with Privacy.tsx / Terms.tsx

Fix classified: reality-sync (factual) vs. escalate (substantive legal wording)

No regulatory compliance assertions in code, commits, or PR text

Self-host vs. cloud context considered if deploy/docs touched

AI data flows stay within disclosed "message the AI Assistant" boundary

Vendor privacy/DPA URL cited — not inferred

Ambiguity flagged to human owner — not resolved by agent alone