agenticapps-observability

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.

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

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.