القائمة
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.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
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.
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.
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.
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.
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.
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.
| name | bondery-skill-authoring |
| description | 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. |
| metadata | {"version":"1.0.0","namespace":"bondery"} |
.agents/skills/description frontmatter for agent discoveryreferences/pnx skills add and updating skills-lock.jsonname must match the directory name — lowercase letters, numbers, hyphens onlydescription must state what the skill does and when to activate it (trigger keywords)SKILL.md under 500 lines — move details to references/metadata.namespace: bonderypnx skills add instead| Type | Location | Naming | Lock file |
|---|---|---|---|
| Bondery first-party | .agents/skills/bondery-{domain}/ | bondery-api, bondery-ux, etc. | No |
| Upstream CLI | .agents/skills/{upstream-name}/ | Keep upstream name (e.g. supabase-postgres-best-practices) | Yes — skills-lock.json |
.agents/skills/bondery-example/
├── SKILL.md # Hub: when to use, non-negotiables, decision tree, checklist
├── references/ # Progressive disclosure (one hop from SKILL.md)
│ ├── README.md # Index table
│ └── topic.md
├── scripts/ # Optional: tested automation agents should run
└── assets/ # Optional: templates, static resources
Minimal frontmatter:
---
name: bondery-example
description: >
One sentence WHAT. Use when [trigger phrases].
metadata:
version: "1.0.0"
namespace: bondery
---
Canonical example: bondery-api — decision tree + merge checklist at end of SKILL.md.
| Task | Read |
|---|---|
| agentskills.io field rules | references/specification.md |
| Context budget, gotchas, defaults | references/best-practices.md |
| Test and iterate on a skill | references/evaluating-skills.md |
| Bundle scripts in a skill | references/scripts.md |
| Bondery naming and checklist pattern | references/bondery-conventions.md |
Full index: references/README.md.
External spec: https://agentskills.io/specification
name matches directory name; lowercase letters, numbers, hyphens onlydescription states WHAT + WHEN + trigger keywords (third person, ≤1024 chars)references/references/ file linked one hop from SKILL.md with "when to read" context../../bondery-api/references/...)pnx skills add ... and commit skills-lock.json — do not hand-edit upstream contentnpx skills-ref validate .agents/skills/<name> (if available).cursor/agents/implementer.md, .cursor/agents/architect.md)