The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it reads and processes untrusted data from repository files containing Git conflict markers.
Ingestion points: Files are identified and read using git diff --name-only --diff-filter=U and git ls-files -u (SKILL.md).
Boundary markers: The skill lacks explicit instructions to treat file content as untrusted data or to ignore instructions embedded within the code, though it directs the agent to analyze surrounding context for intent.
Capability inventory: The skill has the capability to execute shell commands via Git and perform file write operations to resolve conflicts.
Sanitization: No specific sanitization or validation of the content within the conflicted files is described to prevent the agent from acting on instructions found within the source code.
[COMMAND_EXECUTION]: The skill invokes several shell commands to manage the Git state and continue operations.
It utilizes GIT_EDITOR=true git rebase --continue and GIT_EDITOR=true git merge --continue to bypass interactive editors and automate the Git workflow. While these are powerful commands, they are used here within the scope of the skill's intended purpose.

resolve-conflicts