Skills/Glama/agent-eval-policy

agent-eval-policy@0.3.0

CAUTION

Publisher claims: "Enables deterministic routing of worker roles to cost tiers with budget admission, validation of agent responses against contracts and source citations, and deduplication/linting of dossier URLs and markdown." Observed in sandbox: no network, touches nothing outside the project, writes no memory.

Registry GlamaPublisher romanfjose · joined 0 days agoInstalls 0Last detonated 1 h ago
What would fix thisWatchDispute verdict

Is [email protected] safe to install?

KillerSkills ran it in the KillerSkills honeypot sandbox (1 scenario) on 2026-11-06 and scored it 68/100 (Caution). It is not known to be malicious, but it has behaviour or code a security policy should decide on before install. Notable findings: Hidden zero-width characters detected; URL with embedded credentials; Bidirectional text override characters detected. Summary: no network, touches nothing outside the project, writes no memory.

Behavior trace

Honeypot run · 0 s · 1 scenarios · 0 deviations
t+CategoryObserved action
0.0sprocessno install scripts in bundle
0.0sprocessscript backend runs setup only
0.0sprocessscript backend runs setup only
0.0sprocessno install scripts to re-run

Static findings

HIGH
Hidden zero-width characters detectedevals/cases.json:133 · Hidden zero-width characters detected — may conceal contentFix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.
AST07
LOW
Bidirectional override character (text renders in a different order than it is read)evals/cases.json:138 · Bidirectional override character (text renders in a different order than it is read) · informational: not an instruction fileFix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.
AST01
MEDIUM
Outbound HTTP to raw IP addresstest/docs.test.mjs:43 · Outbound HTTP to raw IP address — may bypass domain-based monitoringFix: Replace encoded or assembled strings with plain literals and remove undeclared network calls.
AST07
MEDIUM
Bidirectional text override characters detectedtest/evaluate.test.mjs:54 · Bidirectional text override characters detected — may hide content directionFix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.
AST07
HIGH
URL with embedded credentialsevals/cases.json:63 · URL with embedded credentials — may be phishing or credential leakingFix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.
AST07
MEDIUM
Outbound HTTP to raw IP addresstest/protocol.test.mjs:112 · Outbound HTTP to raw IP address — may bypass domain-based monitoringFix: Replace encoded or assembled strings with plain literals and remove undeclared network calls.
AST07
MEDIUM
npx invocation of non-well-known packagepromptfooconfig.yaml:1 · npx invocation of non-well-known package — may execute untrusted codeFix: Pin versions and sources; vendor anything fetched at install time.
AST06
HIGH
Bidirectional text override characters detectedevals/cases.json:159 · Bidirectional text override characters detected — may hide content directionFix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.
AST07
MEDIUM
URL using raw IP addresstest/docs.test.mjs:43 · URL using raw IP address — may bypass domain-based filteringFix: Replace encoded or assembled strings with plain literals and remove undeclared network calls.
AST07
MEDIUM
Hidden zero-width characters detectedtest/evaluate.test.mjs:51 · Hidden zero-width characters detected — may conceal contentFix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.
AST07
HIGH
Hidden zero-width characters detectedevals/cases.json:136 · Hidden zero-width characters detected — may conceal contentFix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.
AST07
MEDIUM
URL using raw IP addresssrc/http.mjs:76 · URL using raw IP address — may bypass domain-based filteringFix: Replace encoded or assembled strings with plain literals and remove undeclared network calls.
AST07
HIGH
Outbound HTTP to raw IP addresssrc/http.mjs:117 · Outbound HTTP to raw IP address — may bypass domain-based monitoringFix: Replace encoded or assembled strings with plain literals and remove undeclared network calls.
AST07
MEDIUM
Outbound HTTP to raw IP addresstest/protocol.test.mjs:136 · Outbound HTTP to raw IP address — may bypass domain-based monitoringFix: Replace encoded or assembled strings with plain literals and remove undeclared network calls.
AST07
HIGH
Outbound HTTP to raw IP addresssrc/http.mjs:76 · Outbound HTTP to raw IP address — may bypass domain-based monitoringFix: Replace encoded or assembled strings with plain literals and remove undeclared network calls.
AST07
HIGH
Outbound HTTP to raw IP addresssrc/http.mjs:118 · Outbound HTTP to raw IP address — may bypass domain-based monitoringFix: Replace encoded or assembled strings with plain literals and remove undeclared network calls.
AST07
MEDIUM
URL using raw IP addresstest/protocol.test.mjs:136 · URL using raw IP address — may bypass domain-based filteringFix: Replace encoded or assembled strings with plain literals and remove undeclared network calls.
AST07
HIGH
Bidirectional text override characters detectedevals/cases.json:138 · Bidirectional text override characters detected — may hide content directionFix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.
AST07
MEDIUM
URL using raw IP addresssrc/http.mjs:118 · URL using raw IP address — may bypass domain-based filteringFix: Replace encoded or assembled strings with plain literals and remove undeclared network calls.
AST07
HIGH
Hidden zero-width characters detectedevals/cases.json:132 · Hidden zero-width characters detected — may conceal contentFix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.
AST07
MEDIUM
URL using raw IP addresssrc/http.mjs:117 · URL using raw IP address — may bypass domain-based filteringFix: Replace encoded or assembled strings with plain literals and remove undeclared network calls.
AST07
HIGH
Hidden zero-width characters detectedevals/cases.json:137 · Hidden zero-width characters detected — may conceal contentFix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.
AST07
LOW
Bidirectional override character (text renders in a different order than it is read)evals/cases.json:159 · Bidirectional override character (text renders in a different order than it is read) · informational: not an instruction fileFix: Put hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.
AST01
MEDIUM
URL using raw IP addresstest/protocol.test.mjs:112 · URL using raw IP address — may bypass domain-based filteringFix: Replace encoded or assembled strings with plain literals and remove undeclared network calls.
AST07
MEDIUM
Access to sensitive file pathstest/protocol.test.mjs:60 · Access to sensitive file paths — SSH keys, AWS credentials, env files, or system filesFix: Stop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.
AST05

Cost to the agent

Instructions loaded
≈46 tok
1 file · 42 in bundle

Tokens estimated at four characters each. Skills load their instruction text when triggered; the figure is what that costs.

Known vulnerabilities

Vulnerability advisories (OSV) cover npm and PyPI packages; this registry has no advisory database.

What would make this Verified

12 points to 80 · weights: behavior 45%, static 25%, prompt 15%, publisher 15%
  1. Unclaimed publisher≈ +8 on the totalClaim romanfjose: a verified publisher gets +50 on publisher trust. It is free and usually instant with GitHub sign-in.
  2. Hidden zero-width characters detected≈ +4 on the totalPut hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.
  3. URL with embedded credentials≈ +4 on the totalStop enumerating or reading credential locations; take secrets from a declared environment variable the user sets on purpose.
  4. Outbound HTTP to raw IP address≈ +2 on the totalReplace encoded or assembled strings with plain literals and remove undeclared network calls.
  5. Bidirectional text override characters detected≈ +2 on the totalPut hosts, commands and strings in plain source. Decoding at runtime looks identical to malware and costs points on its own.
  6. npx invocation of non-well-known package≈ +2 on the totalPin versions and sources; vendor anything fetched at install time.
Publish a new version to Glama, then queue it here. Not your skill? Claim the handle to get verdict changes by email.Claim romanfjosePropose a patch to the maintainer
agent-eval-policy@0.3.0 safety verdict: Caution 68/100 — KillerSkills