Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Agent Session Cost
v1.0.0Track and analyze your OpenClaw session costs. Parse transcripts, calculate per-model spend, set budgets, alert on overruns. Stop burning money blindly.
⭐ 0· 358·0 current·0 all-time
byFLY@imaflytok
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
Name/description claim to parse OpenClaw session logs and compute spend; the SKILL.md only reads ~/.openclaw/agents/*/sessions/*.jsonl and computes totals and per-model breakdowns, which is exactly what the skill says it will do.
Instruction Scope
Runtime instructions read session files under the user's home directory (~/.openclaw/agents/*/sessions/*.jsonl) which is appropriate for cost analysis. However the SKILL.md also contains an HTML comment with OADP metadata including three external URLs (hub, reg, ping) on onlyflies.buzz — the file does not explain their purpose. That metadata could be inert, but it could also indicate an out-of-band registration or telemetry endpoint not described in the skill, which is scope creep.
Install Mechanism
Instruction-only skill with no install spec and no code files. No downloads or archive extraction are requested, which is low-risk.
Credentials
The skill requests no environment variables, no credentials, and no config paths beyond reading the user's OpenClaw session files (which is required for the stated purpose). There are no declared secrets or unrelated credentials.
Persistence & Privilege
Skill is not always-enabled and does not request elevated or persistent platform privileges. Instructions propose adding a periodic 'heartbeat' entry (user-managed), but the skill itself does not request automatic always-on presence.
Scan Findings in Context
[embedded_remote_urls_in_comment] unexpected: SKILL.md includes an HTML comment: '<!-- OADP:1.0 hub=https://onlyflies.buzz/... reg=https://onlyflies.buzz/... ping=https://onlyflies.buzz/... -->'. The skill does not document contacting these endpoints. This is not necessary for local cost calculation and could indicate a registration/telemetry hook or installer metadata.
What to consider before installing
This skill appears to do what it says: read your local OpenClaw session logs and compute spend using small Python snippets. That behavior is proportionate to the description and requires no credentials. However, the SKILL.md contains an unexplained metadata comment that lists external URLs (onlyflies.buzz). Before installing, consider: (1) ask the publisher what those URLs are for and why they're embedded; (2) remove or sanitize the comment if you only want local-only behavior; (3) run the skill in a sandbox or inspect runtime behavior to confirm it does not send data externally; (4) if you install it, avoid granting extra credentials and monitor network egress from the agent process. If the publisher can confirm the OADP metadata is inert or provides trustworthy documentation for those endpoints, the concern would be resolved. Otherwise treat this as potentially exfiltration-capable telemetry and proceed with caution.Like a lobster shell, security has layers — review code before you run it.
latestvk97fevjzach31jq2pxjassrznh823srb
License
MIT-0
Free to use, modify, and redistribute. No attribution required.