find4— agent skill

find4 — independently scanned and version-tracked by SaferSkills.

by rondomondo·Agent Skill·github.com/rondomondo/agent-skills-beta

Is find4 safe to install?

SaferSkills independently audited find4 (Agent Skill) and scored it 45/100 (orange). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 33 high-severity and 2 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.

Score
45/100
●●●●●○○○○○
↑ +0 since first scan (45 → 45)Re-scan~30s
Latest scan
ScannedJun 28, 2026 · 28d ago
Scans run1 over 90 days
Detectors55 checks · 5 categories
Findings2 warnings · 33 high
EngineSaferSkills 2b638c6
View methodology →
SaferSkills installs
This week0
This month0
All time0
CategoryWeightCategory scoreContribution
Securityprompt, exec, net, exfil, eval
35%
0
0.0 pts
Supply chainhash, typosquat, maintainer, lockfile
20%
100
20.0 pts
Maintenancestaleness, pinning, CI
15%
100
15.0 pts
TransparencySKILL.md, perms, README
15%
100
15.0 pts
Communityinstalls, verify, response
15%
100
15.0 pts

Findings & checks · 35 flagged

Securityscore 0 · 35 findings
HIGHLong base64-encoded blob hidden in the skill documentationSS-SKILL-INJECT-B64-PAYLOAD-01 · Prompt injection · skills/find4/README.md×2
HIGHonce decoded by the agent, an encoded payload has the same impact class as plain-text injection.
Why it matters

A base64 string of 128+ characters appears in a documentation file. Encoded prompt injection hides the hostile instruction in base64 — invisible to keyword filters — and relies on the agent's ability to decode it at runtime. There is no normal authoring reason to embed a multi-hundred-byte base64 blob in skill docs.

The exact value spotted
excerptskills/find4/README.md· markdown
399[img-bpftrace]: https://find4.org/examples/bpftrace-one-liners-and-internals.html.png
400 
401[link-bpftrace]: https://find4.org/#game=1Zh9b9u6FYe_CuELDClmxS-Jk9j_OY7SGnFiI3bveu9WGJRE2Zx
… (108 chars elided on L401)
402 
403[img-brendan-gregg]: https://find4.org/examples/brendan-gregg.png
Occurrences
2 occurrences · first at L401, also L413
Show all 2 locations
Line
File
L401
skills/find4/README.md
L413
skills/find4/README.md
How to fix
Remove the encoded blob, or decode it and review what it actually contains.
  1. Decode the base64 string and confirm it is not an instruction directed at the agent.
  2. Move any legitimate binary or signature data into a dedicated file (*.sig, SIGNATURES) outside the documentation.
Framework references
OWASPLLM01ATLASAML.T0051
Trace & refs
ruleSS-SKILL-INJECT-B64-PAYLOAD-01sha2563a6739e517b5d1ebrubric 365aacaView on GitHub
HIGHLong base64-encoded blob hidden in the skill documentationSS-SKILL-INJECT-B64-PAYLOAD-01 · Prompt injection · skills/find4/games/construction-business-and-site-delivery.html_share_url.json
HIGHonce decoded by the agent, an encoded payload has the same impact class as plain-text injection.
Why it matters

A base64 string of 128+ characters appears in a documentation file. Encoded prompt injection hides the hostile instruction in base64 — invisible to keyword filters — and relies on the agent's ability to decode it at runtime. There is no normal authoring reason to embed a multi-hundred-byte base64 blob in skill docs.

The exact value spotted
excerptskills/find4/games/construction-business-and-site-delivery.html_share_url.json· json
15"ratio": "36%"
16},
17"share_url": "https://find4.org/#game=rVhdd9o4EP0rPrwuDoSEtMmb-WjqUwJZQjbt7snhCHmw1ciWV7KhtC
… (106 chars elided on L17)
18}
Occurrences
1 occurrence · at L17
How to fix
Remove the encoded blob, or decode it and review what it actually contains.
  1. Decode the base64 string and confirm it is not an instruction directed at the agent.
  2. Move any legitimate binary or signature data into a dedicated file (*.sig, SIGNATURES) outside the documentation.
Framework references
OWASPLLM01ATLASAML.T0051
Trace & refs
ruleSS-SKILL-INJECT-B64-PAYLOAD-01sha256678bcafb46cf1b8crubric 365aacaView on GitHub
HIGHLong base64-encoded blob hidden in the skill documentationSS-SKILL-INJECT-B64-PAYLOAD-01 · Prompt injection · skills/find4/games/equine-veterinary-clinical-skills-in-demand.html_share_url.json
HIGHonce decoded by the agent, an encoded payload has the same impact class as plain-text injection.
Why it matters

A base64 string of 128+ characters appears in a documentation file. Encoded prompt injection hides the hostile instruction in base64 — invisible to keyword filters — and relies on the agent's ability to decode it at runtime. There is no normal authoring reason to embed a multi-hundred-byte base64 blob in skill docs.

The exact value spotted
excerptskills/find4/games/equine-veterinary-clinical-skills-in-demand.html_share_url.json· json
15"ratio": "38%"
16},
17"share_url": "https://find4.org/#game=tVhtU-u6Ef4rmnwteSEHwoFvJjjgNsQZ29BLW4ZR5LWjiy35yDI0Pc
… (106 chars elided on L17)
18}
Occurrences
1 occurrence · at L17
How to fix
Remove the encoded blob, or decode it and review what it actually contains.
  1. Decode the base64 string and confirm it is not an instruction directed at the agent.
  2. Move any legitimate binary or signature data into a dedicated file (*.sig, SIGNATURES) outside the documentation.
Framework references
OWASPLLM01ATLASAML.T0051
Trace & refs
ruleSS-SKILL-INJECT-B64-PAYLOAD-01sha256bfa715299ef68a66rubric 365aacaView on GitHub
HIGHLong base64-encoded blob hidden in the skill documentationSS-SKILL-INJECT-B64-PAYLOAD-01 · Prompt injection · skills/find4/games/equine-veterinary-clinical-skills.html_share_url.json
HIGHonce decoded by the agent, an encoded payload has the same impact class as plain-text injection.
Why it matters

A base64 string of 128+ characters appears in a documentation file. Encoded prompt injection hides the hostile instruction in base64 — invisible to keyword filters — and relies on the agent's ability to decode it at runtime. There is no normal authoring reason to embed a multi-hundred-byte base64 blob in skill docs.

The exact value spotted
excerptskills/find4/games/equine-veterinary-clinical-skills.html_share_url.json· json
15"ratio": "36%"
16},
17"share_url": "https://find4.org/#game=pVhtb5tIEP4ryNJ9umD8njjS6UQc0nC1jYVxerlTFK1hjbcBlu6CXV
… (106 chars elided on L17)
18}
Occurrences
1 occurrence · at L17
How to fix
Remove the encoded blob, or decode it and review what it actually contains.
  1. Decode the base64 string and confirm it is not an instruction directed at the agent.
  2. Move any legitimate binary or signature data into a dedicated file (*.sig, SIGNATURES) outside the documentation.
Framework references
OWASPLLM01ATLASAML.T0051
Trace & refs
ruleSS-SKILL-INJECT-B64-PAYLOAD-01sha2566c8e87728fa1f6f1rubric 365aacaView on GitHub
HIGHLong base64-encoded blob hidden in the skill documentationSS-SKILL-INJECT-B64-PAYLOAD-01 · Prompt injection · skills/find4/games/fabulous-irish-mammies-things-she-says.html_share_url.json
HIGHonce decoded by the agent, an encoded payload has the same impact class as plain-text injection.
Why it matters

A base64 string of 128+ characters appears in a documentation file. Encoded prompt injection hides the hostile instruction in base64 — invisible to keyword filters — and relies on the agent's ability to decode it at runtime. There is no normal authoring reason to embed a multi-hundred-byte base64 blob in skill docs.

The exact value spotted
excerptskills/find4/games/fabulous-irish-mammies-things-she-says.html_share_url.json· json
15"ratio": "34%"
16},
17"share_url": "https://find4.org/#game=xZdtb-I4EMe_ipU3fXFAKYWi8uZES9oitbACer3dE0ImmSTeOnbOdu
… (106 chars elided on L17)
18}
Occurrences
1 occurrence · at L17
How to fix
Remove the encoded blob, or decode it and review what it actually contains.
  1. Decode the base64 string and confirm it is not an instruction directed at the agent.
  2. Move any legitimate binary or signature data into a dedicated file (*.sig, SIGNATURES) outside the documentation.
Framework references
OWASPLLM01ATLASAML.T0051
Trace & refs
ruleSS-SKILL-INJECT-B64-PAYLOAD-01sha256ee9f40ae81914ec0rubric 365aacaView on GitHub
HIGHLong base64-encoded blob hidden in the skill documentationSS-SKILL-INJECT-B64-PAYLOAD-01 · Prompt injection · skills/find4/games/famous-bmx-racers-legends-and-pioneers.html_share_url.json
HIGHonce decoded by the agent, an encoded payload has the same impact class as plain-text injection.
Why it matters

A base64 string of 128+ characters appears in a documentation file. Encoded prompt injection hides the hostile instruction in base64 — invisible to keyword filters — and relies on the agent's ability to decode it at runtime. There is no normal authoring reason to embed a multi-hundred-byte base64 blob in skill docs.

The exact value spotted
excerptskills/find4/games/famous-bmx-racers-legends-and-pioneers.html_share_url.json· json
15"ratio": "37%"
16},
17"share_url": "https://find4.org/#game=zVj9byI5Ev1XLE6jS7QQCB9Jlt8a6AQmfEQ0zEzuNkKm24An3W3W7Q
… (106 chars elided on L17)
18}
Occurrences
1 occurrence · at L17
How to fix
Remove the encoded blob, or decode it and review what it actually contains.
  1. Decode the base64 string and confirm it is not an instruction directed at the agent.
  2. Move any legitimate binary or signature data into a dedicated file (*.sig, SIGNATURES) outside the documentation.
Framework references
OWASPLLM01ATLASAML.T0051
Trace & refs
ruleSS-SKILL-INJECT-B64-PAYLOAD-01sha256e737c8028deb03e8rubric 365aacaView on GitHub
HIGHLong base64-encoded blob hidden in the skill documentationSS-SKILL-INJECT-B64-PAYLOAD-01 · Prompt injection · skills/find4/games/famous-fat-people-entertainment-and-culture.html_share_url.json×2
HIGHonce decoded by the agent, an encoded payload has the same impact class as plain-text injection.
Why it matters

A base64 string of 128+ characters appears in a documentation file. Encoded prompt injection hides the hostile instruction in base64 — invisible to keyword filters — and relies on the agent's ability to decode it at runtime. There is no normal authoring reason to embed a multi-hundred-byte base64 blob in skill docs.

The exact value spotted
excerptskills/find4/games/famous-fat-people-entertainment-and-culture.html_share_url.json· json
15"ratio": "37%"
16},
17"share_url": "https://find4.org/#game=tZhtU-M4FoX_iipVsztTGyCEAB2-OWmHuMkLm0CnmF3KpdiKrUG2PL
… (106 chars elided on L17)
18}
Occurrences
2 occurrences · first at L17, also L17
Show all 2 locations
Line
File
L17
skills/find4/games/famous-fat-people-entertainment-and-culture.html_share_url.json
L17
skills/find4/games/famous-fat-people-entertainment-and-culture.html_share_url.json
How to fix
Remove the encoded blob, or decode it and review what it actually contains.
  1. Decode the base64 string and confirm it is not an instruction directed at the agent.
  2. Move any legitimate binary or signature data into a dedicated file (*.sig, SIGNATURES) outside the documentation.
Framework references
OWASPLLM01ATLASAML.T0051
Trace & refs
ruleSS-SKILL-INJECT-B64-PAYLOAD-01sha256a49bc8e9b5735928rubric 365aacaView on GitHub
HIGHLong base64-encoded blob hidden in the skill documentationSS-SKILL-INJECT-B64-PAYLOAD-01 · Prompt injection · skills/find4/games/famous-fat-people-history-and-royalty.html_share_url.json×2
HIGHonce decoded by the agent, an encoded payload has the same impact class as plain-text injection.
Why it matters

A base64 string of 128+ characters appears in a documentation file. Encoded prompt injection hides the hostile instruction in base64 — invisible to keyword filters — and relies on the agent's ability to decode it at runtime. There is no normal authoring reason to embed a multi-hundred-byte base64 blob in skill docs.

The exact value spotted
excerptskills/find4/games/famous-fat-people-history-and-royalty.html_share_url.json· json
15"ratio": "37%"
16},
17"share_url": "https://find4.org/#game=tZjxc9o4Fsf_FQ0ze92dg4aQkqT5zRDANIA5IMtk7zIexRa2NrLlle
… (106 chars elided on L17)
18}
Occurrences
2 occurrences · first at L17, also L17
Show all 2 locations
Line
File
L17
skills/find4/games/famous-fat-people-history-and-royalty.html_share_url.json
L17
skills/find4/games/famous-fat-people-history-and-royalty.html_share_url.json
How to fix
Remove the encoded blob, or decode it and review what it actually contains.
  1. Decode the base64 string and confirm it is not an instruction directed at the agent.
  2. Move any legitimate binary or signature data into a dedicated file (*.sig, SIGNATURES) outside the documentation.
Framework references
OWASPLLM01ATLASAML.T0051
Trace & refs
ruleSS-SKILL-INJECT-B64-PAYLOAD-01sha2560f83fec651325956rubric 365aacaView on GitHub
HIGHLong base64-encoded blob hidden in the skill documentationSS-SKILL-INJECT-B64-PAYLOAD-01 · Prompt injection · skills/find4/games/fastapi-request-handling.html_share_url.json
HIGHonce decoded by the agent, an encoded payload has the same impact class as plain-text injection.
Why it matters

A base64 string of 128+ characters appears in a documentation file. Encoded prompt injection hides the hostile instruction in base64 — invisible to keyword filters — and relies on the agent's ability to decode it at runtime. There is no normal authoring reason to embed a multi-hundred-byte base64 blob in skill docs.

The exact value spotted
excerptskills/find4/games/fastapi-request-handling.html_share_url.json· json
15"ratio": "33%"
16},
17"share_url": "https://find4.org/#game=tZdtc-I2EMe_ioa-LGAgPCR55wMn0BKgtpm7XJthhCWMerLlSnIIuc
… (106 chars elided on L17)
18}
Occurrences
1 occurrence · at L17
How to fix
Remove the encoded blob, or decode it and review what it actually contains.
  1. Decode the base64 string and confirm it is not an instruction directed at the agent.
  2. Move any legitimate binary or signature data into a dedicated file (*.sig, SIGNATURES) outside the documentation.
Framework references
OWASPLLM01ATLASAML.T0051
Trace & refs
ruleSS-SKILL-INJECT-B64-PAYLOAD-01sha256c3b99425b02148d4rubric 365aacaView on GitHub
HIGHLong base64-encoded blob hidden in the skill documentationSS-SKILL-INJECT-B64-PAYLOAD-01 · Prompt injection · skills/find4/games/html/construction-business-and-site-delivery.html_share_url.json
HIGHonce decoded by the agent, an encoded payload has the same impact class as plain-text injection.
Why it matters

A base64 string of 128+ characters appears in a documentation file. Encoded prompt injection hides the hostile instruction in base64 — invisible to keyword filters — and relies on the agent's ability to decode it at runtime. There is no normal authoring reason to embed a multi-hundred-byte base64 blob in skill docs.

The exact value spotted
excerptskills/find4/games/html/construction-business-and-site-delivery.html_share_url.json· json
15"ratio": "36%"
16},
17"share_url": "https://find4.org/#game=rVhdd9o4EP0rPrwuDoSEtMmb-WjqUwJZQjbt7snhCHmw1ciWV7KhtC
… (106 chars elided on L17)
18}
Occurrences
1 occurrence · at L17
How to fix
Remove the encoded blob, or decode it and review what it actually contains.
  1. Decode the base64 string and confirm it is not an instruction directed at the agent.
  2. Move any legitimate binary or signature data into a dedicated file (*.sig, SIGNATURES) outside the documentation.
Framework references
OWASPLLM01ATLASAML.T0051
Trace & refs
ruleSS-SKILL-INJECT-B64-PAYLOAD-01sha256678bcafb46cf1b8crubric 365aacaView on GitHub
HIGHLong base64-encoded blob hidden in the skill documentationSS-SKILL-INJECT-B64-PAYLOAD-01 · Prompt injection · skills/find4/games/html/equine-veterinary-clinical-skills-in-demand.html_share_url.json
HIGHonce decoded by the agent, an encoded payload has the same impact class as plain-text injection.
Why it matters

A base64 string of 128+ characters appears in a documentation file. Encoded prompt injection hides the hostile instruction in base64 — invisible to keyword filters — and relies on the agent's ability to decode it at runtime. There is no normal authoring reason to embed a multi-hundred-byte base64 blob in skill docs.

The exact value spotted
excerptskills/find4/games/html/equine-veterinary-clinical-skills-in-demand.html_share_url.json· json
15"ratio": "38%"
16},
17"share_url": "https://find4.org/#game=tVhtU-u6Ef4rmnwteSEHwoFvJjjgNsQZ29BLW4ZR5LWjiy35yDI0Pc
… (106 chars elided on L17)
18}
Occurrences
1 occurrence · at L17
How to fix
Remove the encoded blob, or decode it and review what it actually contains.
  1. Decode the base64 string and confirm it is not an instruction directed at the agent.
  2. Move any legitimate binary or signature data into a dedicated file (*.sig, SIGNATURES) outside the documentation.
Framework references
OWASPLLM01ATLASAML.T0051
Trace & refs
ruleSS-SKILL-INJECT-B64-PAYLOAD-01sha256bfa715299ef68a66rubric 365aacaView on GitHub
HIGHLong base64-encoded blob hidden in the skill documentationSS-SKILL-INJECT-B64-PAYLOAD-01 · Prompt injection · skills/find4/games/html/equine-veterinary-clinical-skills.html_share_url.json
HIGHonce decoded by the agent, an encoded payload has the same impact class as plain-text injection.
Why it matters

A base64 string of 128+ characters appears in a documentation file. Encoded prompt injection hides the hostile instruction in base64 — invisible to keyword filters — and relies on the agent's ability to decode it at runtime. There is no normal authoring reason to embed a multi-hundred-byte base64 blob in skill docs.

The exact value spotted
excerptskills/find4/games/html/equine-veterinary-clinical-skills.html_share_url.json· json
15"ratio": "36%"
16},
17"share_url": "https://find4.org/#game=pVhtb5tIEP4ryNJ9umD8njjS6UQc0nC1jYVxerlTFK1hjbcBlu6CXV
… (106 chars elided on L17)
18}
Occurrences
1 occurrence · at L17
How to fix
Remove the encoded blob, or decode it and review what it actually contains.
  1. Decode the base64 string and confirm it is not an instruction directed at the agent.
  2. Move any legitimate binary or signature data into a dedicated file (*.sig, SIGNATURES) outside the documentation.
Framework references
OWASPLLM01ATLASAML.T0051
Trace & refs
ruleSS-SKILL-INJECT-B64-PAYLOAD-01sha2566c8e87728fa1f6f1rubric 365aacaView on GitHub
HIGHLong base64-encoded blob hidden in the skill documentationSS-SKILL-INJECT-B64-PAYLOAD-01 · Prompt injection · skills/find4/games/html/fabulous-irish-mammies-things-she-says.html_share_url.json
HIGHonce decoded by the agent, an encoded payload has the same impact class as plain-text injection.
Why it matters

A base64 string of 128+ characters appears in a documentation file. Encoded prompt injection hides the hostile instruction in base64 — invisible to keyword filters — and relies on the agent's ability to decode it at runtime. There is no normal authoring reason to embed a multi-hundred-byte base64 blob in skill docs.

The exact value spotted
excerptskills/find4/games/html/fabulous-irish-mammies-things-she-says.html_share_url.json· json
15"ratio": "34%"
16},
17"share_url": "https://find4.org/#game=xZdtb-I4EMe_ipU3fXFAKYWi8uZES9oitbACer3dE0ImmSTeOnbOdu
… (106 chars elided on L17)
18}
Occurrences
1 occurrence · at L17
How to fix
Remove the encoded blob, or decode it and review what it actually contains.
  1. Decode the base64 string and confirm it is not an instruction directed at the agent.
  2. Move any legitimate binary or signature data into a dedicated file (*.sig, SIGNATURES) outside the documentation.
Framework references
OWASPLLM01ATLASAML.T0051
Trace & refs
ruleSS-SKILL-INJECT-B64-PAYLOAD-01sha256ee9f40ae81914ec0rubric 365aacaView on GitHub
HIGHLong base64-encoded blob hidden in the skill documentationSS-SKILL-INJECT-B64-PAYLOAD-01 · Prompt injection · skills/find4/games/html/famous-bmx-racers-legends-and-pioneers.html_share_url.json
HIGHonce decoded by the agent, an encoded payload has the same impact class as plain-text injection.
Why it matters

A base64 string of 128+ characters appears in a documentation file. Encoded prompt injection hides the hostile instruction in base64 — invisible to keyword filters — and relies on the agent's ability to decode it at runtime. There is no normal authoring reason to embed a multi-hundred-byte base64 blob in skill docs.

The exact value spotted
excerptskills/find4/games/html/famous-bmx-racers-legends-and-pioneers.html_share_url.json· json
15"ratio": "37%"
16},
17"share_url": "https://find4.org/#game=zVj9byI5Ev1XLE6jS7QQCB9Jlt8a6AQmfEQ0zEzuNkKm24An3W3W7Q
… (106 chars elided on L17)
18}
Occurrences
1 occurrence · at L17
How to fix
Remove the encoded blob, or decode it and review what it actually contains.
  1. Decode the base64 string and confirm it is not an instruction directed at the agent.
  2. Move any legitimate binary or signature data into a dedicated file (*.sig, SIGNATURES) outside the documentation.
Framework references
OWASPLLM01ATLASAML.T0051
Trace & refs
ruleSS-SKILL-INJECT-B64-PAYLOAD-01sha256e737c8028deb03e8rubric 365aacaView on GitHub
HIGHLong base64-encoded blob hidden in the skill documentationSS-SKILL-INJECT-B64-PAYLOAD-01 · Prompt injection · skills/find4/games/html/famous-fat-people-entertainment-and-culture.html_share_url.json×2
HIGHonce decoded by the agent, an encoded payload has the same impact class as plain-text injection.
Why it matters

A base64 string of 128+ characters appears in a documentation file. Encoded prompt injection hides the hostile instruction in base64 — invisible to keyword filters — and relies on the agent's ability to decode it at runtime. There is no normal authoring reason to embed a multi-hundred-byte base64 blob in skill docs.

The exact value spotted
excerptskills/find4/games/html/famous-fat-people-entertainment-and-culture.html_share_url.json· json
15"ratio": "37%"
16},
17"share_url": "https://find4.org/#game=tZhtU-M4FoX_iipVsztTGyCEAB2-OWmHuMkLm0CnmF3KpdiKrUG2PL
… (106 chars elided on L17)
18}
Occurrences
2 occurrences · first at L17, also L17
Show all 2 locations
Line
File
L17
skills/find4/games/html/famous-fat-people-entertainment-and-culture.html_share_url.json
L17
skills/find4/games/html/famous-fat-people-entertainment-and-culture.html_share_url.json
How to fix
Remove the encoded blob, or decode it and review what it actually contains.
  1. Decode the base64 string and confirm it is not an instruction directed at the agent.
  2. Move any legitimate binary or signature data into a dedicated file (*.sig, SIGNATURES) outside the documentation.
Framework references
OWASPLLM01ATLASAML.T0051
Trace & refs
ruleSS-SKILL-INJECT-B64-PAYLOAD-01sha256a49bc8e9b5735928rubric 365aacaView on GitHub
HIGHLong base64-encoded blob hidden in the skill documentationSS-SKILL-INJECT-B64-PAYLOAD-01 · Prompt injection · skills/find4/games/html/famous-fat-people-history-and-royalty.html_share_url.json×2
HIGHonce decoded by the agent, an encoded payload has the same impact class as plain-text injection.
Why it matters

A base64 string of 128+ characters appears in a documentation file. Encoded prompt injection hides the hostile instruction in base64 — invisible to keyword filters — and relies on the agent's ability to decode it at runtime. There is no normal authoring reason to embed a multi-hundred-byte base64 blob in skill docs.

The exact value spotted
excerptskills/find4/games/html/famous-fat-people-history-and-royalty.html_share_url.json· json
15"ratio": "37%"
16},
17"share_url": "https://find4.org/#game=tZjxc9o4Fsf_FQ0ze92dg4aQkqT5zRDANIA5IMtk7zIexRa2NrLlle
… (106 chars elided on L17)
18}
Occurrences
2 occurrences · first at L17, also L17
Show all 2 locations
Line
File
L17
skills/find4/games/html/famous-fat-people-history-and-royalty.html_share_url.json
L17
skills/find4/games/html/famous-fat-people-history-and-royalty.html_share_url.json
How to fix
Remove the encoded blob, or decode it and review what it actually contains.
  1. Decode the base64 string and confirm it is not an instruction directed at the agent.
  2. Move any legitimate binary or signature data into a dedicated file (*.sig, SIGNATURES) outside the documentation.
Framework references
OWASPLLM01ATLASAML.T0051
Trace & refs
ruleSS-SKILL-INJECT-B64-PAYLOAD-01sha2560f83fec651325956rubric 365aacaView on GitHub
HIGHLong base64-encoded blob hidden in the skill documentationSS-SKILL-INJECT-B64-PAYLOAD-01 · Prompt injection · skills/find4/games/html/fastapi-request-handling.html_share_url.json
HIGHonce decoded by the agent, an encoded payload has the same impact class as plain-text injection.
Why it matters

A base64 string of 128+ characters appears in a documentation file. Encoded prompt injection hides the hostile instruction in base64 — invisible to keyword filters — and relies on the agent's ability to decode it at runtime. There is no normal authoring reason to embed a multi-hundred-byte base64 blob in skill docs.

The exact value spotted
excerptskills/find4/games/html/fastapi-request-handling.html_share_url.json· json
15"ratio": "33%"
16},
17"share_url": "https://find4.org/#game=tZdtc-I2EMe_ioa-LGAgPCR55wMn0BKgtpm7XJthhCWMerLlSnIIuc
… (106 chars elided on L17)
18}
Occurrences
1 occurrence · at L17
How to fix
Remove the encoded blob, or decode it and review what it actually contains.
  1. Decode the base64 string and confirm it is not an instruction directed at the agent.
  2. Move any legitimate binary or signature data into a dedicated file (*.sig, SIGNATURES) outside the documentation.
Framework references
OWASPLLM01ATLASAML.T0051
Trace & refs
ruleSS-SKILL-INJECT-B64-PAYLOAD-01sha256c3b99425b02148d4rubric 365aacaView on GitHub
HIGHLong base64-encoded blob hidden in the skill documentationSS-SKILL-INJECT-B64-PAYLOAD-01 · Prompt injection · skills/find4/games/html/landing-a-great-sre-job-career-strategy-and-culture-fit.html_share_url.json×2
HIGHonce decoded by the agent, an encoded payload has the same impact class as plain-text injection.
Why it matters

A base64 string of 128+ characters appears in a documentation file. Encoded prompt injection hides the hostile instruction in base64 — invisible to keyword filters — and relies on the agent's ability to decode it at runtime. There is no normal authoring reason to embed a multi-hundred-byte base64 blob in skill docs.

The exact value spotted
excerptskills/find4/games/html/landing-a-great-sre-job-career-strategy-and-culture-fit.html_share_url.json· json
15"ratio": "39%"
16},
17"share_url": "https://find4.org/#game=vVhdc-I6Ev0rKl7uw2IgJJl8vHnAEN8hOIXJZLO7qZSwhVFiLF9JDs
… (106 chars elided on L17)
18}
Occurrences
2 occurrences · first at L17, also L17
Show all 2 locations
Line
File
L17
skills/find4/games/html/landing-a-great-sre-job-career-strategy-and-culture-fit.html_share_url.json
L17
skills/find4/games/html/landing-a-great-sre-job-career-strategy-and-culture-fit.html_share_url.json
How to fix
Remove the encoded blob, or decode it and review what it actually contains.
  1. Decode the base64 string and confirm it is not an instruction directed at the agent.
  2. Move any legitimate binary or signature data into a dedicated file (*.sig, SIGNATURES) outside the documentation.
Framework references
OWASPLLM01ATLASAML.T0051
Trace & refs
ruleSS-SKILL-INJECT-B64-PAYLOAD-01sha2568537ef9811ad4af2rubric 365aacaView on GitHub
HIGHLong base64-encoded blob hidden in the skill documentationSS-SKILL-INJECT-B64-PAYLOAD-01 · Prompt injection · skills/find4/games/html/space-telescopes-and-black-hole-anatomy.html_share_url.json
HIGHonce decoded by the agent, an encoded payload has the same impact class as plain-text injection.
Why it matters

A base64 string of 128+ characters appears in a documentation file. Encoded prompt injection hides the hostile instruction in base64 — invisible to keyword filters — and relies on the agent's ability to decode it at runtime. There is no normal authoring reason to embed a multi-hundred-byte base64 blob in skill docs.

The exact value spotted
excerptskills/find4/games/html/space-telescopes-and-black-hole-anatomy.html_share_url.json· json
15"ratio": "38%"
16},
17"share_url": "https://find4.org/#game=rVhtc9q6Ev4rGr6cmXNjDKQQkm8OuOCWAteQ9LxMhhGysHVjSz6SDC
… (106 chars elided on L17)
18}
Occurrences
1 occurrence · at L17
How to fix
Remove the encoded blob, or decode it and review what it actually contains.
  1. Decode the base64 string and confirm it is not an instruction directed at the agent.
  2. Move any legitimate binary or signature data into a dedicated file (*.sig, SIGNATURES) outside the documentation.
Framework references
OWASPLLM01ATLASAML.T0051
Trace & refs
ruleSS-SKILL-INJECT-B64-PAYLOAD-01sha25667ae3d05b8cc10b6rubric 365aacaView on GitHub
HIGHLong base64-encoded blob hidden in the skill documentationSS-SKILL-INJECT-B64-PAYLOAD-01 · Prompt injection · skills/find4/games/html/the-irish-mammy-phrase-book.html_share_url.json
HIGHonce decoded by the agent, an encoded payload has the same impact class as plain-text injection.
Why it matters

A base64 string of 128+ characters appears in a documentation file. Encoded prompt injection hides the hostile instruction in base64 — invisible to keyword filters — and relies on the agent's ability to decode it at runtime. There is no normal authoring reason to embed a multi-hundred-byte base64 blob in skill docs.

The exact value spotted
excerptskills/find4/games/html/the-irish-mammy-phrase-book.html_share_url.json· json
15"ratio": "38%"
16},
17"share_url": "https://find4.org/#game=vZdvb9o6FMa_isWbvbilo10LLe8opG3uKKAkbbVdVcjEJnhN7FzboW
… (106 chars elided on L17)
18}
Occurrences
1 occurrence · at L17
How to fix
Remove the encoded blob, or decode it and review what it actually contains.
  1. Decode the base64 string and confirm it is not an instruction directed at the agent.
  2. Move any legitimate binary or signature data into a dedicated file (*.sig, SIGNATURES) outside the documentation.
Framework references
OWASPLLM01ATLASAML.T0051
Trace & refs
ruleSS-SKILL-INJECT-B64-PAYLOAD-01sha25663bbf75090b86716rubric 365aacaView on GitHub
HIGHLong base64-encoded blob hidden in the skill documentationSS-SKILL-INJECT-B64-PAYLOAD-01 · Prompt injection · skills/find4/games/landing-a-great-sre-job-career-strategy-and-culture-fit.html_share_url.json×2
HIGHonce decoded by the agent, an encoded payload has the same impact class as plain-text injection.
Why it matters

A base64 string of 128+ characters appears in a documentation file. Encoded prompt injection hides the hostile instruction in base64 — invisible to keyword filters — and relies on the agent's ability to decode it at runtime. There is no normal authoring reason to embed a multi-hundred-byte base64 blob in skill docs.

The exact value spotted
excerptskills/find4/games/landing-a-great-sre-job-career-strategy-and-culture-fit.html_share_url.json· json
15"ratio": "39%"
16},
17"share_url": "https://find4.org/#game=vVhdc-I6Ev0rKl7uw2IgJJl8vHnAEN8hOIXJZLO7qZSwhVFiLF9JDs
… (106 chars elided on L17)
18}
Occurrences
2 occurrences · first at L17, also L17
Show all 2 locations
Line
File
L17
skills/find4/games/landing-a-great-sre-job-career-strategy-and-culture-fit.html_share_url.json
L17
skills/find4/games/landing-a-great-sre-job-career-strategy-and-culture-fit.html_share_url.json
How to fix
Remove the encoded blob, or decode it and review what it actually contains.
  1. Decode the base64 string and confirm it is not an instruction directed at the agent.
  2. Move any legitimate binary or signature data into a dedicated file (*.sig, SIGNATURES) outside the documentation.
Framework references
OWASPLLM01ATLASAML.T0051
Trace & refs
ruleSS-SKILL-INJECT-B64-PAYLOAD-01sha2568537ef9811ad4af2rubric 365aacaView on GitHub
HIGHLong base64-encoded blob hidden in the skill documentationSS-SKILL-INJECT-B64-PAYLOAD-01 · Prompt injection · skills/find4/games/space-telescopes-and-black-hole-anatomy.html_share_url.json
HIGHonce decoded by the agent, an encoded payload has the same impact class as plain-text injection.
Why it matters

A base64 string of 128+ characters appears in a documentation file. Encoded prompt injection hides the hostile instruction in base64 — invisible to keyword filters — and relies on the agent's ability to decode it at runtime. There is no normal authoring reason to embed a multi-hundred-byte base64 blob in skill docs.

The exact value spotted
excerptskills/find4/games/space-telescopes-and-black-hole-anatomy.html_share_url.json· json
15"ratio": "38%"
16},
17"share_url": "https://find4.org/#game=rVhtc9q6Ev4rGr6cmXNjDKQQkm8OuOCWAteQ9LxMhhGysHVjSz6SDC
… (106 chars elided on L17)
18}
Occurrences
1 occurrence · at L17
How to fix
Remove the encoded blob, or decode it and review what it actually contains.
  1. Decode the base64 string and confirm it is not an instruction directed at the agent.
  2. Move any legitimate binary or signature data into a dedicated file (*.sig, SIGNATURES) outside the documentation.
Framework references
OWASPLLM01ATLASAML.T0051
Trace & refs
ruleSS-SKILL-INJECT-B64-PAYLOAD-01sha25667ae3d05b8cc10b6rubric 365aacaView on GitHub
HIGHLong base64-encoded blob hidden in the skill documentationSS-SKILL-INJECT-B64-PAYLOAD-01 · Prompt injection · skills/find4/games/the-irish-mammy-phrase-book.html_share_url.json
HIGHonce decoded by the agent, an encoded payload has the same impact class as plain-text injection.
Why it matters

A base64 string of 128+ characters appears in a documentation file. Encoded prompt injection hides the hostile instruction in base64 — invisible to keyword filters — and relies on the agent's ability to decode it at runtime. There is no normal authoring reason to embed a multi-hundred-byte base64 blob in skill docs.

The exact value spotted
excerptskills/find4/games/the-irish-mammy-phrase-book.html_share_url.json· json
15"ratio": "38%"
16},
17"share_url": "https://find4.org/#game=vZdvb9o6FMa_isWbvbilo10LLe8opG3uKKAkbbVdVcjEJnhN7FzboW
… (106 chars elided on L17)
18}
Occurrences
1 occurrence · at L17
How to fix
Remove the encoded blob, or decode it and review what it actually contains.
  1. Decode the base64 string and confirm it is not an instruction directed at the agent.
  2. Move any legitimate binary or signature data into a dedicated file (*.sig, SIGNATURES) outside the documentation.
Framework references
OWASPLLM01ATLASAML.T0051
Trace & refs
ruleSS-SKILL-INJECT-B64-PAYLOAD-01sha25663bbf75090b86716rubric 365aacaView on GitHub
HIGHLong base64-encoded blob hidden in the skill documentationSS-SKILL-INJECT-B64-PAYLOAD-01 · Prompt injection · README.md×3
HIGHonce decoded by the agent, an encoded payload has the same impact class as plain-text injection.
Why it matters

A base64 string of 128+ characters appears in a documentation file. Encoded prompt injection hides the hostile instruction in base64 — invisible to keyword filters — and relies on the agent's ability to decode it at runtime. There is no normal authoring reason to embed a multi-hundred-byte base64 blob in skill docs.

The exact value spotted
excerptREADME.md· markdown
326 
327[img-bpftrace]: skills/find4/examples/bpftrace-one-liners-and-internals.html.png
328[link-bpftrace]: https://find4.org/#game=1Zh9b9u6FYe_CuELDClmxS-Jk9j_OY7SGnFiI3bveu9WGJRE2Zx
… (108 chars elided on L328)
329 
330[img-brendan-gregg]: skills/find4/examples/brendan-gregg.png
Occurrences
3 occurrences · first at L328, also L340, L340
Show all 3 locations
Line
File
L328
README.md
L340
README.md
L340
README.md
How to fix
Remove the encoded blob, or decode it and review what it actually contains.
  1. Decode the base64 string and confirm it is not an instruction directed at the agent.
  2. Move any legitimate binary or signature data into a dedicated file (*.sig, SIGNATURES) outside the documentation.
Framework references
OWASPLLM01ATLASAML.T0051
Trace & refs
ruleSS-SKILL-INJECT-B64-PAYLOAD-01sha2563a6739e517b5d1ebrubric 365aacaView on GitHub
MEDIUMInstruction telling the agent not to ask for approvalSS-SKILL-INJECT-DONT-ASK-01 · Prompt injection · skills/find4/SKILL.md×2
MEDIUMit fires on intent; the real damage depends on the host agent's own approval-gating.
Why it matters

The text {match} tells the agent to skip the normal "ask the user first" gate. Used adversarially it removes the human-in-the-loop check before destructive or sensitive actions, turning a normally-gated agent into a fire-and-forget executor.

The exact value spotted
excerptskills/find4/SKILL.md· markdown
107 
108Once the passphrase is in hand (or GitHub push is confirmed skipped), proceed through all re
… (59 chars elided on L108)
109Do NOT ask clarifying questions mid-run unless a hard error requires user input.
110 
111---
Occurrences
2 occurrences · first at L109, also L115
Show all 2 locations
Line
File
L109
skills/find4/SKILL.md
L115
skills/find4/SKILL.md
How to fix
Remove the approval-skipping instruction, or scope it narrowly to a specific safe, reversible action.
  1. Delete blanket "don't ask / no need to confirm" directives from the skill.
  2. If the skill is a genuine autonomous job, restrict the opt-out to a named non-destructive action rather than all actions.
Framework references
OWASPLLM01ATLASAML.T0051
Trace & refs
ruleSS-SKILL-INJECT-DONT-ASK-01sha2567d1c5f93d8975babrubric 365aacaView on GitHub
Supply chainscore 100 · 0 findings
All supply chain checks passedNo findings in this category for the latest scan.pass
Maintenancescore 100 · 0 findings
All maintenance checks passedNo findings in this category for the latest scan.pass
Transparencyscore 100 · 0 findings
All transparency checks passedNo findings in this category for the latest scan.pass
Communityscore 100 · 0 findings
All community checks passedNo findings in this category for the latest scan.pass
Vendor response · right of reply
Are you the maintainer? Submit a response →

Audit the pieces. Scan the whole. Decide.

~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.