Install
$ agentstack add skill-konh-claudetools-implement ✓ 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.
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
Implement the plan at docs/specs/_/plan.md.
Plan Discovery
With $ARGUMENTS: resolve against docs/specs/ — look for docs/specs/_/plan.md using the full dated spec identifier.
Without $ARGUMENTS:
- List all
*/plan.mdfiles underdocs/specs/ - Extract the
YY_MM_DD_HHprefix from each folder - Use the file with the latest timestamp prefix
Orchestration
Read the plan file first. Then decide based on plan size:
- 1–2 steps or trivial changes: implement inline (no sub-agent needed)
- 3+ steps, or steps spanning distinct areas of the codebase: spawn a developer sub-agent per major phase
When spawning a developer sub-agent, brief it with:
- The full plan text
- Which step(s) it is responsible for
- Relevant project rules (
CLAUDE.md, any project rules directory) — code style and any domain-specific conventions - Current file state / context it needs (read key files first and include excerpts)
- Whether any external tool the plan depends on (an editor, engine, or service with its own connection state) is available and connected
- For steps touching source code: write the test for the new behavior first so it fails against the current code, then implement until it passes — never disable or weaken an existing test to force a pass, fix the underlying code instead
After each sub-agent phase:
- Relay its results and any errors to the user
- Verify the work (check for errors, read changed files) before moving to the next phase
- If a phase fails, diagnose and re-brief the sub-agent with the fix context — do not skip ahead
- Mark each completed agent step in the plan file by changing
- [ ]to- [x]
Pre-flight Checks
- Follow all project standards in
CLAUDE.mdand any project rules directory - If the plan depends on an external tool or editor with its own connection/session state, verify that connection is live before starting. If it isn't, stop and ask the user to establish it first.
Completion
After all steps are done:
- If any changes touch source code: confirm tests were written test-first (red before green) for the affected logic and all pass
- Run the
code-reviewskill (or/code-reviewif unavailable as a skill) on the changed files — present any concerns one by one and ask the user to approve each fix before applying it
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: KonH
- Source: KonH/ClaudeTools
- 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.