The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to perform project operations using the algokit CLI for initialization, building, and testing.
Evidence: Usage of algokit init, algokit project run build, and algokit project run test in SKILL.md.
[EXTERNAL_DOWNLOADS]: The skill fetches templates and examples from official Algorand Foundation public GitHub repositories.
Evidence: References to algorandfoundation/devportal-code-examples and algorandfoundation/puya in SKILL.md.
[INDIRECT_PROMPT_INJECTION]: The skill facilitates the development and testing of user-provided smart contract logic, which represents an attack surface for untrusted data ingestion.
Ingestion points: User-supplied Python contract code and transaction parameters in SKILL.md and references/testing.md.
Boundary markers: Absent; instructions do not define specific delimiters for isolating untrusted user code from agent instructions.
Capability inventory: Shell command execution via algokit and local file system operations in SKILL.md.
Sanitization: Absent; no explicit validation steps are provided for user-inputted contract source code.
[DYNAMIC_EXECUTION]: The development workflow involves the runtime compilation of Python source code into TEAL bytecode using the PuyaPy compiler.
Evidence: Compilation step using algokit project run build for contract generation in SKILL.md and references/deploy-interaction.md.

algorand-python