The Agent Skills Directory

[SAFE]: No malicious code, prompt injections, or unauthorized activities were detected. The skill is focused on legitimate Amazon seller operations and adheres to the author's vendor infrastructure.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the Amazon marketplace (competitor offer sets).
Ingestion points: External data is ingested through getItemOffers, getCompetitivePricing, and ANY_OFFER_CHANGED notifications as described in references/01-sp-api-surfaces.md.
Boundary markers: The skill explicitly separates data retrieval from execution by requiring a human-reviewable 'change-set' as defined in references/05-report-template.md.
Capability inventory: The skill has read access to competitive data and write access to price attributes via the patchListingsItem operation.
Sanitization: The skill mandates the use of mode=VALIDATION_PREVIEW to surface pricing policy violations and Featured Offer disqualification before any changes are applied. This human-in-the-loop requirement effectively mitigates the risk of the agent acting on malicious payloads embedded in offer data.
[COMMAND_EXECUTION]: Tool usage is strictly limited to authorized SP-API operations. There are no instructions to execute arbitrary shell commands, access sensitive local configuration files, or perform privilege escalation.

amazon-sp-repricing