Issue Station
Sign In
Tools hereToolbox

openclaw - 💡(How to fix) Fix [Feature]: Automatic post-response skill execution / agent_end hook [1 participants]

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…
GitHub stats
openclaw/openclaw#54357•Fetched 2026-04-08 01:28:34
View on GitHub
Comments
0
Participants
1
Timeline
1
Reactions
0
Author
cindulasaicindulasai
Participants
cindulasaicindulasai
Timeline (top)
labeled ×1

OpenClaw skills are instruction-driven (model reads SKILL.md and follows rules), not automated. Nextsteps only fires
if I remember — and I missed some replies in this thread. Tessl.io/VSCode has a post-processing layer that
automatically injects nextsteps on every response, which is why it works there.

But the same skill doesn't work here in OpenClaw TUI or in telegram

https://clawhub.ai/cindulasai/nextsteps https://tessl.io/registry/crab/nextsteps

https://github.com/cindulasai/nextsteps

Root Cause

Skills that need to run after every response (like nextsteps) are unreliable because they depend on the agent
remembering, not enforced automation.

RAW_BUFFERClick to expand / collapse

Summary

OpenClaw skills are instruction-driven (model reads SKILL.md and follows rules), not automated. Nextsteps only fires
if I remember — and I missed some replies in this thread. Tessl.io/VSCode has a post-processing layer that
automatically injects nextsteps on every response, which is why it works there.

But the same skill doesn't work here in OpenClaw TUI or in telegram

https://clawhub.ai/cindulasai/nextsteps https://tessl.io/registry/crab/nextsteps

https://github.com/cindulasai/nextsteps

Problem to solve

Skills that need to run after every response (like nextsteps) are unreliable because they depend on the agent
remembering, not enforced automation.

Proposed solution

  1. Add agent_end plugin hook — fires after every response, lets a runner auto-inject nextsteps
  2. Add auto: true flag in SKILL.md — framework auto-executes skills marked automatic after each response
  3. Document the limitation until one of the above is implemented

Alternatives considered

No response

Impact

  • User expectation gap — Users who install nextsteps from ClawHub expect it to work like Tessl.io/VSCode —
  •       https://tessl.io/registry/crab/nextsteps

automatically after every message. Currently it doesn't, creating confusion and disappointment.

  • Skill ecosystem — Many quality-of-life skills depend on per-response execution (auto-memory snapshots, smart
    reminders, response tagging, tone checkers). Without an automation layer, all of them are unreliable by design.
  • Adoption — ClawHub skills that require automation are effectively broken out of the box. This hurts the skill
    ecosystem and OpenClaw's credibility as a platform.
  • Token overhead — The instruction-driven approach wastes tokens every turn (re-reading SKILL.md rules, pipeline
    steps). An automated hook would be more efficient.

Severity: Medium — doesn't break core functionality, but undermines a whole class of skill-based features.

Evidence/examples

No response

Additional information

No response

extent analysis

Fix Plan

To address the issue of skills not running automatically after every response, we will implement the proposed solution of adding an agent_end plugin hook and an auto: true flag in SKILL.md. Here are the steps:

  • Step 1: Add agent_end plugin hook
    • Create a new plugin that listens for the agent_end event
    • In the event handler, check if the response requires auto-injection of nextsteps
    • If required, inject the nextsteps into the response
  • Step 2: Add auto: true flag in SKILL.md
    • Update the SKILL.md file to include an auto: true flag for skills that require automation
    • Create a framework to auto-execute skills marked as automatic after each response

Example code for the agent_end plugin hook:

import os

def agent_end_plugin(agent, response):
    # Check if the response requires auto-injection of nextsteps
    if response.requires_nextsteps:
        # Inject nextsteps into the response
        response.nextsteps = get_nextsteps(response)

def get_nextsteps(response):
    # Logic to get nextsteps based on the response
    pass

Example code for the auto: true flag in SKILL.md:

name: nextsteps
auto: true

Verification

To verify that the fix worked, test the skills that require automation after every response. Check if the nextsteps are injected correctly and if the skills are executed as expected.

Extra Tips

  • Document the limitation of the current implementation until the automation layer is implemented
  • Consider optimizing the token overhead by reducing the number of times the SKILL.md rules are re-read
  • Ensure that the automation layer is efficient and does not introduce any performance issues

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…
#logging issue#authentication issue#prompt issue#agent setup#task chaining

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING

openclaw - 💡(How to fix) Fix [Feature]: Automatic post-response skill execution / agent_end hook [1 participants]