New skills: read-runtime-manager-logs — fetch application logs across CH1/CH2/RTF

[tbolis-at-mulesoft]

Problem

Reading runtime application logs is one of the most common MuleSoft operator/developer tasks, but today an AI agent using this repo has to hand-compose calls across three different APIs depending on where the app runs:

• CloudHub 1.0 → cloudhub (/v2/applications/{domain}/logs, /instances/{instanceId}/logs, /organizations/{orgid}/logs/download)
• CloudHub 2.0 → cloudhub-20 (/organizations/{org}/privatespaces/{psId}/logs)
• Runtime Fabric / AMC → amc-application-manager (.../deployments/{deploymentId}/specs/{specId}/logs + /logs/file)

There is no JTBD that orchestrates this. The closest existing skill is anypoint-platform-operations, which doesn't cover log retrieval. The result is that agents either pick the wrong endpoint for the deployment target or fail to chain the prerequisite lookups (org → env → deployment → spec).

Proposed skill

Add a new skill `skills/read-runtime-manager-logs/SKILL.md` that:

1. Detects the deployment target (CH1 / CH2 / RTF) from the caller's context or a prior lookup.
2. Resolves the required identifiers (org, env, deployment/app, instance or spec).
3. Fetches logs using the correct API for that target.
4. Returns a normalized result the agent can summarize or grep.

Happy to open a PR once the scope is confirmed and the `enhancement` label is applied.