Skip to main content
在 Manus 中运行任何 Skill
一键导入
agenticapps-eu
GitHub 创作者资料

agenticapps-eu

按仓库查看 5 个 GitHub 仓库中的 35 个已收集 skills。

已收集 skills
35
仓库
5
更新
2026-07-14
仓库浏览

仓库与代表性 skills

agentic-apps-workflow
其他计算机职业

Enforces the AgenticApps spec-first workflow on opencode. This skill MUST activate whenever the user asks opencode to implement, build, code, fix, refactor, or design anything. Triggers on: "let's work on [issue]", "implement the [feature]", "build the [component]", "fix the [bug]", any task involving writing or changing code, creating architecture, or making technical decisions. Use this even when the user just says "start working" or references a Linear / Asana / Jira / GitHub issue number. The skill emits the workflow commitment ritual, picks task size, routes to the right GSD entry-point skill, and binds every spec gate to the opencode-* gate-fulfilling skill that satisfies it.

2026-07-14
setup-opencode-agenticapps-workflow
软件开发工程师

Bootstrap a fresh project with the opencode-workflow scaffolding by installing the LATEST snapshot directly — no migration replay. Lays down the current end-state project-side artifacts (AGENTS.md sections, .planning/config.json, .opencode/workflow-config.md, docs/decisions/, .opencode/workflow-version.txt) in one shot and stamps the current scaffolder version. Use when a project is freshly cloned or initialized and the user asks to "set up the workflow", "add agenticapps workflow", "enable opencode-workflow", "install the discipline layer", "scaffold this project", or anything else that means "I want this project to use opencode-workflow from this point forward". Idempotent — refuses to re-run on a project that already has `.opencode/workflow-version.txt` and routes to `$update-opencode-agenticapps-workflow` instead.

2026-07-14
opencode-database-sentinel-audit
信息安全分析师

Database-specialist security audit in two modes. **phase-scoped mode**: when the security gate fires AND the phase scope matches supabase|postgres|mongodb|firebase|mysql, audit the changed surface for RLS / auth bypass / storage exposure / known CVEs (MongoBleed, pgBouncer, mysql_native_password drift) and produce DB-AUDIT.md. **pre-launch mode**: before first production launch and after any major DB migration, audit the FULL project surface (every supported backend). Critical or High findings BLOCK branch close per ADR-0012; override only via the database-security acceptance ADR template.

2026-07-01
opencode-impeccable-audit
网页与数字界面设计师

Post-implementation visual quality audit against the deployed component. Re-scores the seven design dimensions (typography, color, spatial, motion, interaction, responsive, UX writing) against the shipping code, runs the 24-item AI-slop anti-pattern scan, and blocks branch close on Red findings (per ADR-0011). Use after a UI-shipping phase's implementation is complete and the changes are visible in the dev server. May also be invoked retroactively against shipped UI to catch drift between mockup and implementation.

2026-07-01
opencode-qa
软件质量保证分析师与测试员

Browser-driven QA in two modes. **ui-preview mode** (per-task, pre-commit): boot the dev server, screenshot the changed component, reference the screenshot path in the commit message — fires for any task modifying a frontend component, route, or visual surface. **qa mode** (post-phase): walk user flows in a real browser against a running dev server, log interactions, produce a QA report referenced from VERIFICATION.md — fires per phase that ships user-visible behavior. Use whichever mode the calling context selects; the body branches on `mode=preview` vs `mode=phase-qa`.

2026-07-01
opencode-spec-review
软件开发工程师

Stage 1 of the two-stage review: audit the phase's changeset against CONTEXT.md decisions, every must_have in VERIFICATION.md, the gate bindings, and the protocol-violation list (commitment ritual, brainstorm artifact, TDD commit pair). Use after all execution tasks complete and before any code-review pass. Writes the `## Stage 1 — Spec compliance` section into REVIEW.md. The implementer agent runs this stage in the same session; Stage 2 (`superpowers:requesting-code-review`) MUST run independently afterwards.

2026-07-01
opencode-ts-declare-first
软件开发工程师

Declare-first TypeScript discipline for new modules in TS-primary AgenticApps projects on the opencode host, per workflow-core spec §13. Strengthens the `tdd` gate for TypeScript module API surfaces. When invoked on a new TS module, produces three ATOMIC commits in order: Phase 1 declaration (`<name>.declare.ts`, declare-only, type-checks clean), Phase 2 tests (import + exercise the declared surface; MUST be observable as failing in the expected way — the RED output is §06 evidence), Phase 3 implementation (signatures match the declaration exactly; tests go green). Use when authoring a new TS module's public API, or when the user says "declare first", "ts-declare-first", "type surface first", or starts a new TypeScript module in a TS-primary project. The skill REFUSES to bundle Phase 1 and Phase 3 into a single commit — the three-commit atomicity is the structural evidence that the discipline was followed; collapsing it erases the evidence.

2026-07-01
opencode-cso
信息安全分析师

Run an OWASP-aligned security audit at a phase boundary when the changeset touches authentication, storage, request handling, secret material, or LLM trust boundaries. Produces SECURITY.md with audited threat models and mitigation evidence. Use when the phase's diff includes auth flows, session handling, file/blob storage, API surfaces, environment variables/secrets, or LLM tool-use trust boundaries. Composes with `opencode-database-sentinel-audit` (DB-specific sub-gate) per ADR-0012.

2026-06-30
当前展示该仓库 Top 8 / 11 个已收集 skills。
agentic-apps-workflow
其他计算机职业

Enforces the AgenticApps spec-first workflow on Codex. This skill MUST activate whenever the user asks Codex to implement, build, code, fix, refactor, or design anything. Triggers on: "let's work on [issue]", "implement the [feature]", "build the [component]", "fix the [bug]", any task involving writing or changing code, creating architecture, or making technical decisions. Use this even when the user just says "start working" or references a Linear / Asana / Jira / GitHub issue number. The skill emits the workflow commitment ritual, picks task size, routes to the right GSD entry-point skill, and binds every spec gate to the upstream Superpowers skill or codex-* gate-fulfilling skill that satisfies it.

2026-07-06
setup-codex-agenticapps-workflow
其他计算机职业

Bootstrap a fresh project with the codex-workflow scaffolding — apply the baseline migration to install the trigger skill's required project-side artifacts (AGENTS.md sections, .planning/config.codex.json, .codex/workflow-config.md, docs/decisions/, .codex/workflow-version.txt). Use when a project is freshly cloned or initialized and the user asks to "set up the workflow", "add agenticapps workflow", "enable codex-workflow", "install the discipline layer", "scaffold this project", or anything else that means "I want this project to use codex-workflow from this point forward". Idempotent — refuses to re-run on a project that already has `.codex/workflow-version.txt` and routes to `$update-codex-agenticapps-workflow` instead.

2026-07-06
update-codex-agenticapps-workflow
软件开发工程师

Update an installed codex-workflow project to the current scaffolder version by applying pending migrations between the project's recorded version and the scaffolder's. Use when the user runs "update workflow", "apply pending workflow migrations", "upgrade to the new scaffolder version", or after pulling a new release of codex-workflow. Reads `.codex/workflow-version.txt`, finds pending migrations in `${CODEX_HOME}/skills/update-codex-agenticapps-workflow/migrations/`, pre-flights required external skills, applies each step with diff preview + per-step confirm, bumps the version on success, commits atomically. Supports `--dry-run` for diff-only output.

2026-07-03
codex-ts-declare-first
软件开发工程师

Declare-first TypeScript discipline for new modules in TS-primary AgenticApps projects on the Codex host, per workflow-core spec §13. Strengthens the `tdd` gate for TypeScript module API surfaces. When invoked on a new TS module, produces three ATOMIC commits in order: Phase 1 declaration (`<name>.declare.ts`, declare-only, type-checks clean), Phase 2 tests (import + exercise the declared surface; MUST be observable as failing in the expected way — the RED output is §06 evidence), Phase 3 implementation (signatures match the declaration exactly; tests go green). Use when authoring a new TS module's public API, or when the user says "declare first", "ts-declare-first", "type surface first", or starts a new TypeScript module in a TS-primary project. The skill REFUSES to bundle Phase 1 and Phase 3 into a single commit — the three-commit atomicity is the structural evidence that the discipline was followed; collapsing it erases the evidence.

2026-07-01
codex-database-sentinel-audit
信息安全分析师

Database-specialist security audit in two modes. **phase-scoped mode**: when the security gate fires AND the phase scope matches supabase|postgres|mongodb|firebase|mysql, audit the changed surface for RLS / auth bypass / storage exposure / known CVEs (MongoBleed, pgBouncer, mysql_native_password drift) and produce DB-AUDIT.md. **pre-launch mode**: before first production launch and after any major DB migration, audit the FULL project surface (every supported backend). Critical or High findings BLOCK branch close per ADR-0012; override only via the database-security acceptance ADR template.

2026-07-01
codex-impeccable-audit
网页与数字界面设计师

Post-implementation visual quality audit against the deployed component. Re-scores the seven design dimensions (typography, color, spatial, motion, interaction, responsive, UX writing) against the shipping code, runs the 24-item AI-slop anti-pattern scan, and blocks branch close on Red findings (per ADR-0011). Use after a UI-shipping phase's implementation is complete and the changes are visible in the dev server. May also be invoked retroactively against shipped UI to catch drift between mockup and implementation.

2026-07-01
codex-qa
软件质量保证分析师与测试员

Browser-driven QA in two modes. **ui-preview mode** (per-task, pre-commit): boot the dev server, screenshot the changed component, reference the screenshot path in the commit message — fires for any task modifying a frontend component, route, or visual surface. **qa mode** (post-phase): walk user flows in a real browser against a running dev server, log interactions, produce a QA report referenced from VERIFICATION.md — fires per phase that ships user-visible behavior. Use whichever mode the calling context selects; the body branches on `mode=preview` vs `mode=phase-qa`.

2026-07-01
codex-spec-review
软件开发工程师

Stage 1 of the two-stage review: audit the phase's changeset against CONTEXT.md decisions, every must_have in VERIFICATION.md, the gate bindings, and the protocol-violation list (commitment ritual, brainstorm artifact, TDD commit pair). Use after all execution tasks complete and before any code-review pass. Writes the `## Stage 1 — Spec compliance` section into REVIEW.md. The implementer agent runs this stage in the same session; Stage 2 (`superpowers:requesting-code-review`) MUST run independently afterwards.

2026-07-01
当前展示该仓库 Top 8 / 11 个已收集 skills。
setup-agenticapps-workflow
软件开发工程师

Bootstrap a fresh project with the AgenticApps Superpowers + GSD + gstack workflow by installing the LATEST snapshot directly — no migration replay. Lays down the current end-state project artifacts (the workflow skill, CLAUDE.md sections, hooks, settings, planning config, version stamp) in one shot. Use when the user runs "/setup-agenticapps-workflow", "set up the workflow", "install agenticapps workflow", or "scaffold this project". Idempotent — refuses to re-run on a project that already has `.claude/skills/agentic-apps-workflow/` and routes to `/update-agenticapps-workflow` instead. Fail-closed on an unverified snapshot: refuses to install until the drift guard (`check-snapshot-parity.sh`) passes, so a raw/stale seed can never be laid down silently.

2026-07-06
agentic-apps-workflow
软件开发工程师

Enforces the spec-first development workflow using Superpowers + GSD + gstack for any AgenticApps project. This skill MUST activate whenever Claude is asked to implement, build, code, fix, refactor, or design anything in the current project — regardless of whether the user explicitly mentions the workflow. Triggers on: "let's work on [issue]", "implement the [feature]", "build the [component]", "fix the [bug]", any task involving writing or changing code, creating architecture, or making technical decisions. The skill ensures every piece of work follows the Superpowers + GSD + gstack discipline and produces traceable decision artefacts. Use this even when the user just says "start working" or references a Linear issue number.

2026-07-06
ts-declare-first
软件开发工程师

Declare-first TypeScript discipline for new modules in TS-primary AgenticApps projects, per workflow-core spec §13. When invoked on a new TS module, produces three ATOMIC commits in order: Phase 1 — declaration: a `declare`-only type-surface file (`<name>.declare.ts`). No implementation bodies, no expression initialisers, no executable code. Type- checks clean against `tsc --noEmit`. Phase 2 — tests: test files that import and exercise the declared surface. Tests MUST be observable as failing in the expected way at this commit (declarations exist, implementations don't). The expected-failure runner output is §06 evidence. Phase 3 — implementation: signatures match the declaration exactly. Tests now pass; commit captures the transition. The skill REFUSES to bundle Phase 1 and Phase 3 into a single commit. The three-commit atomicity is the structural evidence that the discipline was followed; coll

2026-05-21
agentic-apps-workflow
软件开发工程师

synthetic test fixture for migration 0009 — before-fresh

2026-05-14
agentic-apps-workflow
软件质量保证分析师与测试员

synthetic test fixture for migration 0009 — before-inlined-customised

2026-05-14
agentic-apps-workflow
软件质量保证分析师与测试员

synthetic test fixture for migration 0009 — before-inlined-pristine

2026-05-14
agentic-apps-workflow
软件质量保证分析师与测试员

synthetic test fixture for migration 0009 — after-idempotent

2026-05-13
agentic-apps-workflow
软件质量保证分析师与测试员

synthetic test fixture for migration 0009 — after-vendored

2026-05-13
当前展示该仓库 Top 8 / 9 个已收集 skills。
observability
软件开发工程师

Generate or audit an observability wrapper that satisfies AgenticApps core spec §10 for the project's tech stack. Three subcommands: `init` (greenfield scaffold — detect stacks, materialise wrapper + middleware + policy behind consent gates, wire the entry file, write the `observability:` metadata block, smoke-verify), `scan` (brownfield validate → .scan-report.md; flags `--since-commit <ref>` for §10.9 delta scans and `--update-baseline` for `.observability/baseline.json`), and `scan-apply` (apply high-confidence findings with per-file consent, regenerating the baseline). Use when adding or auditing logging, tracing, error capture, spans, or Sentry/OTLP/Axiom destinations, or when wiring a §10 observability wrapper into a new or existing service. Per-stack templates and the full per-subcommand procedure live in this skill's body and ./init, ./scan, ./scan-apply, ./templates. Host-neutral — /observability on Claude, $observability on Codex.

2026-06-17
injection-guard
信息安全分析师

Generate or audit a prompt-injection defense that satisfies AgenticApps core spec §14 for the project's tech stack. Three subcommands: `init` (greenfield scaffold — detect stacks, materialise the trust-classification enforcement (TS ESLint rule / Go analyzer), the `fenceUntrusted` + canary helpers, the untrusted-input registry, and the attack-matrix regression test behind consent gates, then write the `injection_guard:` metadata block), `scan` (brownfield audit → .injection-scan-report.md; finds tenant-untrusted values reaching prompts without a trust marker, ranked by confidence), and `scan-apply` (apply high-confidence findings — add markers, scaffold missing enforcement assets — with per-file consent). Use when adding or auditing an LLM prompt-building path, tenant-untrusted tagging, prompt fencing, output validation / canary checks, tool-dispatch allowlists, or a §14 prompt-injection regression test in a new or existing service. The shared attack-family corpus and per-stack templates live in ./corpus, ./t

2026-06-17
add-observability
软件开发工程师

Deprecated alias for the `observability` skill — renamed in 0.11.0. Use `observability` (`/observability` on Claude, `$observability` on Codex) instead. Retained for the 0.11.0–0.12.0 alias window; removed in 0.14.0. Routes init / scan / scan-apply to the canonical skill. This field is required by stricter skill loaders (e.g. Codex 0.130.0, which deep-scans nested SKILL.md files); without it the loader errors on this alias.

2026-06-09
已展示 5 / 5 个仓库
已展示全部仓库