skill-gb3h-agent-skills-executing-plans · by Install
$ agentstack add skill-gb3h-agent-skills-executing-plans ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Links to your public security report.[](https://agentstack.voostack.com/security/report/skill-gb3h-agent-skills-executing-plans) Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
Executing Plans
Load plan. Review critically. Execute all tasks. Verify each step. Report when complete.
If subagents are available: consider using subagent-driven-development instead — fresh context per task yields higher quality.
Step 1: Load and Review Plan
- Read the plan file with
read - Review critically — identify questions, concerns, missing context, or unclear steps
- If concerns exist: raise them with the user BEFORE starting. Do not silently work around plan gaps.
- If no concerns: proceed to execution
Step 2: Execute Tasks
For each task in the plan:
- Read the step — understand exactly what it asks
- Execute exactly what the step says — no scope creep, no "while I'm here" improvements
- Verify with concrete evidence (file exists, test passes, command succeeds). Use
execandread— not assumptions. - Mark done — update the plan:
[x] - Continue to next step
When to Stop and Ask for Help
STOP executing immediately when:
- Step fails and you can't diagnose why within 2 attempts
- Plan has a gap preventing the next step
- You don't understand an instruction
- Verification fails repeatedly
- Something contradicts the plan's assumptions
Ask for clarification rather than guessing. Guessing compounds errors across subsequent steps.
When to Revisit Earlier Steps
- User updates the plan based on your feedback
- A later step reveals an earlier step was done incorrectly
- Fundamental approach needs rethinking
Don't force through blockers. Stop, surface the issue, wait for direction.
Rules
- Never start on main/master branch without explicit user consent
- Stay on task. Unrelated issues go in
memory/YYYY-MM-DD.md— don't fix them mid-step - No silent skips. If you skip a step, state why and the consequences
- Use
systematic-debuggingif a step fails unexpectedly - Use
verification-before-completionbefore declaring any step done
Checkpointing
At checkpoint steps (or every 3-5 tasks), write to memory/YYYY-MM-DD.md:
- Steps completed so far
- Current state of work
- Decisions made during execution
- Exact next step to resume from
Resuming from Checkpoint
- Read the most recent memory file with
read - Find the last checkpoint
- Verify checkpoint state is still valid (files exist, no conflicts)
- Resume from the next step
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: gb3h
- Source: gb3h/agent-skills
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.