Engrm Sentinel

Use Engrm packs and Sentinel context to surface likely mistakes, risky patterns, and lessons that would have prevented them.

Security Scan

The skill is an instruction-only reviewer that asks the agent to use existing Engrm context if present; its requirements and instructions are coherent with its stated purpose and it does not request extra privileges or install code.

Purpose & Capability

Name and description claim to surface risky patterns using Engrm packs; the SKILL.md consistently instructs the agent to consult Engrm only if it is already available and to limit actions to review-style guidance. No unrelated credentials, binaries, or installs are requested.

Instruction Scope

Runtime instructions are narrow and reviewer-focused: ask sentinel-style questions, reuse pack lessons, and avoid inventing Engrm commands or performing setup. The skill does not instruct reading arbitrary system files, sending data to external endpoints, or collecting secrets.

Install Mechanism

No install spec and no code files — lowest-risk format. Nothing will be written to disk or downloaded by the skill itself.

Credentials

The skill declares no environment variables or credentials and does not require config paths. It only depends on Engrm being present in the environment, which the instructions explicitly treat as pre-existing and optional.

Persistence & Privilege

always is false and there is no installation step that modifies agent configuration or other skills. Autonomous invocation is allowed (platform default) but this skill does not request elevated persistence or cross-skill configuration changes.

Assessment

This skill is instruction-only and appears coherent and low-risk: it expects Engrm (packs/context) to already be connected and otherwise will not perform setup or access secrets. Before installing, confirm whether your environment actually provides Engrm and what permissions or data Engrm packs expose (packs themselves could reference project-specific content). If you do not have Engrm integrated, the skill will do nothing — it will not attempt to install or contact external endpoints. If you rely on autonomous agent actions, remember the agent could apply this guidance without prompting, so ensure you trust the Engrm packs available in your workspace.

Use this skill when code is entering a risky area and the goal is to prevent mistakes, not just remember past work.

Before you start

Use Engrm only if it is already connected and available in the current environment.

If Engrm is not available, say that Engrm Sentinel context is not connected on this machine and continue without inventing setup steps or shell commands.

Command guardrails

Do not invent Engrm shell commands like engrm search, engrm save, or engrm timeline.

Treat Sentinel guidance as Engrm workflow knowledge, not as an excuse to guess at extra CLI syntax.

What this skill is for

• Surface risky patterns before they turn into bugs or security issues.
• Reuse pack knowledge and standards during active coding work.
• Explain what is going wrong, not just that something looks unusual.
• Keep prevention practical and grounded in project context.

When to use it

Use this skill when the task involves:

• authentication or secrets
• input handling or API boundaries
• security-sensitive data
• major refactors
• deployment or config changes
• code that has caused repeated issues before

Sentinel mindset

Ask:

• What could go wrong here?
• Which pack lessons are relevant?
• Is this repeating a known mistake?
• Is the implementation drifting into a risky shortcut?
• Would a future session benefit from a new standard or lesson here?

Good outputs

Helpful Sentinel guidance should:

• name the likely mistake clearly
• explain why it matters
• connect it to a reusable lesson or pack
• suggest the safer path

Avoid:

• generic fearmongering
• abstract policy language without code relevance
• blocking everything just because it feels risky

Pack-aware review

Remember:

• packs are available on free plans too
• the point is not to hide lessons behind paid access
• the real value is turning those lessons into smarter review, stronger prevention, and better follow-through

What success looks like

The agent catches meaningful mistakes early, explains them well, and leaves behind a sharper memory or standard for the next session.