Home / Articles / Agent hooks in October 2026 / Versions / v1.2.0
Agent hooks in October 2026: what changed in v1.2.0
From v1.1.0 (2026-10-14, 2d0143c83) to v1.2.0 (2026-10-27, db67c5ce1), paragraph by paragraph.
0 paragraphs added, 0 removed, 2 changed in place, 46 unchanged. About 0 words added and 79 removed. Insertions are marked like this, deletions like this; unchanged runs are folded to one line; figures appear as their file names.
2 unchanged paragraphs
!shot hq-map.webp | images/ | Agent hooks in October 2026, from each vendor's documentation:documentation, whethercompared aon hookblocking, can block a tool call, rewrite its input, act at the model call, be forced by the organisation, fail closed,rewriting and whetherfailing hooks shipped in a repository need trust. The design has converged; the security properties have not.closed. | comparison
10 unchanged paragraphs, under In short {#in-short}, What a hook is {#what}
!shot hq-where.webp | images/ | Where a hook sits: beside the agent, on the same machine, with the user's permissions.permissions, Hooks fire when a prompt is submitted, before and after a tool call,firing at thefive stop, and on one platform around the model call itself. They are good at deterministic checks at the moment of action, and cannot undo what ran or see what was not routed through them.moments. | architecture
34 unchanged paragraphs, under The lay of the land {#platforms}, Where hooks fail {#fail}, A small test of my own {#test}…