session-watchdog
Monitor Stalled Claude Sessions and Resume Pending Work
Idle agent sessions can conceal unfinished tasks and unresolved blockers. Session Watchdog reviews recent session activity, requests targeted continuation, and records wake history and heartbeat status.
Install with my Agent
Copy this request to your Agent. It includes the canonical Skill page and manifest.
Review the Skillstore skill "session-watchdog" from https://skillstore.io/skills/rabbitdesigns22-session-watchdog.md and its manifest at https://skillstore.io/api/skills/rabbitdesigns22-session-watchdog/manifest. Verify the artifact. You may proceed after verification, subject to the environment's own policy.Your Agent should still show its plan and request any confirmation required by the security policy.
Agent-readable resources
Use these links when an AI agent, crawler, or script needs clean context instead of reading the full page.
Test it
Using "session-watchdog". Review my sessions without waking any.
Expected outcome:
- Illustrative review: API tests appear stopped mid-task after 28 idle minutes.
- Data export needs your login confirmation.
- Documentation session appears complete; no action proposed.
Using "session-watchdog". Check whether the watchdog heartbeat is overdue.
Expected outcome:
Illustrative result: The last heartbeat is six hours old. The application may have closed or the scheduled run may have stalled.
Using "session-watchdog". Check whether a session can be woken again.
Expected outcome:
Illustrative result: Wake denied. This session was woken 90 minutes ago; the minimum interval is 180 minutes.
Security Audit
Medium RiskMost static alerts are Markdown delimiters, documented permission settings, or expected local logging rather than malicious behavior. The time command executes an unaudited sibling skill, and separate quota checks and recording permit concurrent wake limits to be exceeded. No evidence found of credential theft, network exfiltration, or instructions to override this audit.
Confirmed security concerns (1)
Capability review items (1)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
Risk Factors
โ๏ธ External commands (19)
๐ Filesystem access (3)
Share & cite this report
Share the versioned assessment report, neutral badge, embed card, and citations. Skillstore reports evidence without deciding whether this Skill is safe.
Copy report link
https://skillstore.io/skills/rabbitdesigns22-session-watchdog/audits/1?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_reportMarkdown badge
[](https://skillstore.io/skills/rabbitdesigns22-session-watchdog?utm_source=security_passport_badge)HTML badge
<a href="https://skillstore.io/skills/rabbitdesigns22-session-watchdog?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/rabbitdesigns22-session-watchdog/security.svg" alt="Skillstore security assessment" loading="lazy"></a>Embed card
<iframe src="https://skillstore.io/embed/skills/rabbitdesigns22-session-watchdog.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>Academic citations (APA ยท BibTeX ยท CFF)
APA citation
rabbitdesigns22. (2026). session-watchdog security audit report (audit version 1) [Author version unspecified]. Skillstore. https://skillstore.io/skills/rabbitdesigns22-session-watchdog/audits/1BibTeX citation
@techreport{rabbitdesigns22-rabbitdesigns22-session-watchdog-2026,
author = {rabbitdesigns22},
title = {session-watchdog security audit report (audit version 1)},
institution = {Skillstore},
year = {2026},
number = {1},
url = {https://skillstore.io/skills/rabbitdesigns22-session-watchdog/audits/1},
note = {Author version unspecified}
}CITATION.cff
cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "session-watchdog security audit report (audit version 1)"
version: "unspecified"
type: report
authors:
- name: "rabbitdesigns22"
date-released: "2026-11-11"
url: "https://skillstore.io/skills/rabbitdesigns22-session-watchdog/audits/1"
identifiers:
- type: other
value: "skillstore:rabbitdesigns22-session-watchdog:audit:1"
description: "Skillstore immutable audit report identifier"
Skillstore Score
Why this score Evidence Confidence: MediumWhat You Can Build
Recover Interrupted Development Tasks
Review recent coding sessions and request continuation when a task stopped after an error or an unfinished promised step.
Identify Blocked Background Work
Distinguish sessions waiting for background tasks from sessions needing a login, approval, or other user input.
Check Scheduled Monitoring Health
Use heartbeat and setup timestamps to detect overdue watchdog runs after the application closes or a scheduled run stalls.
Try These Prompts
Review recent sessions with session-watchdog. Do not send wake messages. Report sessions that appear stalled and identify anything requiring my input.
Review sessions active within 24 hours. Check wake history before messaging. Wake at most five eligible sessions and leave completed or user-blocked sessions untouched.
List today's wake records and check whether the heartbeat is over five hours old. Do not wake sessions or change scheduling.
Inspect scheduler support and exact tool permissions. Propose a two-hour watchdog schedule and heartbeat checks. Wait for approval before changing settings or sending messages.
Best Practices
- Review a classification-only run before enabling messaging, scheduling, or prompt hooks.
- Use narrowly scoped permissions and review the optional external time helper before permitting its execution.
- Keep one watchdog run active at a time, check wake history, and verify each scheduled run leaves a heartbeat.
Avoid
- Treating every idle session as unfinished or waking sessions that require user approval.
- Running overlapping watchdogs and assuming separate quota checks guarantee hard wake limits.
- Granting broad shell permissions or trusting scheduler success labels without checking the final result and heartbeat.
Frequently Asked Questions
Does this skill start its own scheduler?
Can I use it with Codex alone?
Which sessions should remain untouched?
Where does it store monitoring data?
Are the wake limits guaranteed?
What happens when a heartbeat becomes overdue?
Developer Details
Author
rabbitdesigns22License
Unknown
Skillstore revision
r1
Version notice
The author did not declare a version.
Ref
f38ac41a6255e57978dd6f276aec539d5e9b5255
Maintenance freshness
11/12/2026
Usage
0 downloads ยท 0 views
File structure