CategoryWeightCategory scoreContribution

Securityprompt, exec, net, exfil, eval

35%

26.6 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 · 2 flagged

Securityscore 76 · 2 findings

MEDIUMInstruction telling the agent not to ask for approvalSS-SKILL-INJECT-DONT-ASK-01 · Prompt injection · skills/technical-readiness/SKILL.md×2

MEDIUM — it 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/technical-readiness/SKILL.md· markdown

233. Inspect accessible project files if access already exists.

244. If more access is needed, ask plainly: "May I inspect this folder/file and tell you what

… (13 chars elided on L24)

255. Do not ask the designer for utility names or technical package names if you can avoid it.

266. Ask only what affects the next step: desired output, result use, whether an export/screen

… (26 chars elided on L26)

277. Suggest work approaches: agent inspection, system app, familiar editor, external service,

… (31 chars elided on L27)

Occurrences

2 occurrences · first at L25, also L35

Show all 2 locations

Line

File

L25

skills/technical-readiness/SKILL.md

L35

skills/technical-readiness/SKILL.md

How to fix

Remove the approval-skipping instruction, or scope it narrowly to a specific safe, reversible action.

Delete blanket "don't ask / no need to confirm" directives from the skill.
If the skill is a genuine autonomous job, restrict the opt-out to a named non-destructive action rather than all actions.

Framework references

OWASPLLM01 ATLASAML.T0051

Trace & refs

ruleSS-SKILL-INJECT-DONT-ASK-01sha25668b030f55bcd484a…rubric 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

The primary manifest — the file an agent reads to learn what this artifact does.

skills/technical-readiness/SKILL.md · 1 file

skills/technical-readiness/SKILL.md2.0 KB · Markdown

name: technical-readiness

description: Use at the start of a copied landscape project folder or before working with photos, PDFs, video, audio, OCR, compression, presentations, web research, or pricing. The agent should infer technical options from accessible files and environment, ask the designer mainly about desired outcomes and permission to inspect files, and avoid scripts or mandatory commands.

Technical Readiness

Goal

Help the agent understand project technical context without scripts, mandatory commands, or shifting diagnosis to the designer.

When To Use

A new designer starts working with a copied template.
The task involves photos, video, PDF, audio, presentation, or prices.
The user asks how to process files.
The agent is about to suggest file processing and must first check whether it is realistic.

Process

Read TECHNICAL_READINESS.md.
Identify file types and desired output.
Inspect accessible project files if access already exists.
If more access is needed, ask plainly: "May I inspect this folder/file and tell you what can be done?"
Do not ask the designer for utility names or technical package names if you can avoid it.
Ask only what affects the next step: desired output, result use, whether an export/screenshot/link can be provided.
Suggest work approaches: agent inspection, system app, familiar editor, external service, local tool, manual workaround.
Record the chosen approach or limitation in 99_project_logs/WORK_LOG.md.

Rules

Do not create scripts/tools automation inside the template.
Do not treat Linux/WSL as the user's default environment.
Do not promise OCR, compression, transcription, or conversion until an approach is agreed.
Do not ask the designer to do technical diagnosis that the agent can infer.
If work can continue without processing the file, continue and mark the gap.

Output

Short conclusion: what can be done now.
Files/exports needed from the designer.
Suitable work approaches.
Recorded limitations.

technical-readiness— agent skill

Is technical-readiness safe to install?

Technical Readiness

Goal

When To Use

Process

Rules

Output

Audit the pieces. Scan the whole. Decide.