ai-job-search released v1.7.3 on October 19. The GitHub release describes 71 entries, including 58 bug fixes closed in two weeks. It also changes the project’s permission configuration for its skills and agents.
The release removes broad tool grants from skill frontmatter, adds deny rules for uploads through curl, wget, scp, ssh, and nc, and restricts reads of .env and credential directories. Agents that process job postings must now declare their tools; the release says posting-fetcher, posting-scorer, and application-reviewer can read, fetch, and search but cannot edit or write.
A new tools/posting_guard.py strips hidden HTML, comments, zero-width and bidi characters, quotes instruction-shaped lines, and flags suspicious hidden URLs or mixed-script words before /apply, /rank, and /interview use a posting. The release says the guard advises rather than blocks. It also fixes salary-data conversion and invalid portal-CLI input handling. The source does not establish independent security testing, exploitability, or whether the new controls prevent prompt injection in all cases.