CtrlK
BlogDocsLog inGet started
Tessl Logo

phi-prompt-guard

Runtime, prompt-time behavioral guardrail that helps reduce PHI exposure in LLM-assisted workflows by detecting PHI-bearing prompts, avoiding unsafe tool actions that would pull more PHI in, and redirecting users toward de-identified or synthetic inputs. Use when the user is about to paste, query, or read clinical/patient data, or when an action (DB query, file read, tool output) may pull PHI into the conversation. Honors a [PHI-OK] attestation for synthetic / test data.

66

Quality

80%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./scientific-skills/Other/phi-prompt-guard/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

67%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A thorough, well-structured behavioral guardrail with concrete reply templates and decision rules. Its main weakness is repetition of the [PHI-OK] and generate-don't-execute guidance across multiple sections, which inflates length without adding clarity.

Suggestions

Consolidate the [PHI-OK] attestation and 'generate, don't execute' rules so each is stated fully once; have the Workflow and Instructions reference the dedicated section instead of restating it.

Move the full 18-identifier list and override-signal heuristics into a reference file (e.g. references/phi-identifiers.md) linked from SKILL.md to reduce body length, provided the in-context decision still works.

Trim the Key Features and Implementation Details sections, which largely restate points already covered in Workflow and Instructions.

DimensionReasoningScore

Conciseness

Mostly efficient, but the [PHI-OK] attestation and 'generate, don't execute' guidance are restated across Workflow, Key Features, Implementation Details, and Instructions, adding repetition that could be consolidated.

3 / 5

Actionability

Concrete reply templates, specific safe-inspection patterns (head -1, df.dtypes, wc -l), and worked SQL examples make the guidance actionable; as an instruction-only skill it lacks runnable code, which is the minor gap.

4 / 5

Workflow Clarity

The Workflow is a clear numbered decision sequence (assess -> proceed -> honor [PHI-OK] -> redirect -> generate-don't-execute -> safer alternative) with explicit detection and override checkpoints; minor redundancy slightly muddies the canonical path.

4 / 5

Progressive Disclosure

Well-organized into clear sections with no nested references, appropriate for a pure in-context behavioral skill where everything must stay in context; the ~174-line single file with some repeated guidance is the minor organization gap.

4 / 5

Total

15

/

20

Passed

Description

92%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong, specific description that concretely states both capability and trigger conditions with a clear clinical-PHI niche. Slight room to broaden natural-language synonyms, but otherwise excellent.

DimensionReasoningScore

Specificity

Lists multiple concrete actions — 'detecting PHI-bearing prompts', 'avoiding unsafe tool actions that would pull more PHI in', 'redirecting users toward de-identified or synthetic inputs' — with comprehensive coverage of what the guardrail does.

5 / 5

Completeness

Explicitly answers both what (detect / avoid / redirect) and when ('Use when the user is about to paste, query, or read clinical/patient data...') with concrete trigger phrases.

5 / 5

Trigger Term Quality

Good natural-keyword coverage ('paste, query, or read clinical/patient data', 'DB query, file read, tool output', 'synthetic / test data'), though the terms lean clinical/technical and a few lay synonyms are absent.

4 / 5

Distinctiveness Conflict Risk

A clearly defined PHI/clinical-data niche with distinctive triggers ([PHI-OK], clinical/patient data, PHI-pulling actions) gives minimal overlap with other skills.

5 / 5

Total

19

/

20

Passed

Validation

93%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

frontmatter_unknown_keys

Unknown frontmatter key(s) found; consider removing or moving to metadata

Warning

Total

15

/

16

Passed

Repository
aipoch/medical-research-skills
Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.