BD pleading-injection-guard
Detects prompt injection hidden in documents from the other side (pleadings, skeletons, bundles, served evidence, opponents' emails and attachments) before an AI reads them, so the model is not turned against your client. Checks word by word that the text an extractor hands the model was visibly drawn on the page: white, near-white, covered, clipped, off-page, tiny, transparent and invisible-mode text in PDF (ink test plus OCR); Word formatting resolved as Word does (styles, shading, theme colours, scaling, off-page frames, fallbacks, unused headers); HTML/CSS, RTF, email with attachments and zips, xlsx/pptx and legacy formats. Flags Unicode smuggling, text addressed to an AI and authorities found only in hidden text. Fails closed. Optional Read-tool hook. Use BEFORE any AI summary or analysis of an opposing or third-party document. Triggers: 'scan for injection', 'is this document safe', 'check the other side's skeleton', 'injection guard'.
Detects prompt injection hidden in documents from the other side (pleadings, skeletons, bundles, served evidence, opponents' emails and attachments) before an…
As a process D 47/100 · Unfinished process — weak spots: result and completion, when it triggers, inputs and preconditions
How to improve
- Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
- A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.
Guard findings · 0
✓ No critical or high findings
Files scanned: 15. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- note
edit-residuethe text marks something as outdated (lines 25, 38, 84): check that old rules are not kept next to new ones — the full check reads the text for contradictions
Process rating: all ten parameters 47/100
- 0Result and completion. Does not say what the result is
- 0Inputs and preconditions. Does not say what the process needs to start
- 0Progress reporting. Says nothing while it works
- 20When it triggers. No condition that starts the skill
- 30Running it twice. 6 mutating operations with no state check
- 40Consistency. Frontmatter name (pleading-injection-guard) differs from the folder (pleading-injection-guard-matei-clej)
- 60Tools and files. Uses tools (bash, web, python) that frontmatter does not declare
- 70Execution cost. Instruction body is 4361 tokens
- 100Steps. 30 steps
- 100Failures and branches. 2 branches, has a failure section
- medium Safety rules and hard prohibitions inside a skill: they belong in the system prompt, here they protect nothing
- low The response is described with custom markup (6 tags): a typed call is more reliable
Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.
Quality signals
- +5Description has no quoted example phrases that should trigger the skill
- +4Description does not say when NOT to use the skill (false activations)
- +3Description length 955: 120–800 characters recommended
- +3Output format is not stated: the model decides each time
- +2Single-language instructions
- +4Structure: 10 headings
- +3Step-by-step instructions: 30 items
- +4Has examples (1 code blocks)
- +1License stated
Quality base 70; lint remarks subtract, signals add up to 100. Result: 82.