Commit Graph

24 Commits

Author SHA1 Message Date
UnaiAlias 1c79507140
CAI 1.1.5 release (#455) 2026-06-05 13:43:19 +02:00
Víctor Mayoral Vilches bfb1281096 Prompt injection mitigation (#249)
* Reproduce prompt injection flaw in CAI

Signed-off-by: Víctor Mayoral Vilches <v.mayoralv@gmail.com>

* Mitigate prompt injection

Signed-off-by: Víctor Mayoral Vilches <v.mayoralv@gmail.com>

* Add a second PoC, refactor and propose a message when guardrails are triggered

Signed-off-by: Víctor Mayoral Vilches <v.mayoralv@gmail.com>

* Add CAI_GUARDRAILS and 2 additional PoCs

Signed-off-by: Víctor Mayoral Vilches <v.mayoralv@gmail.com>

* Mitigate PoC 3

Signed-off-by: Víctor Mayoral Vilches <v.mayoralv@gmail.com>

* Mitigate PoC 4

Signed-off-by: Víctor Mayoral Vilches <v.mayoralv@gmail.com>

* Add and mitigate PoCs 5 and 6

Signed-off-by: Víctor Mayoral Vilches <v.mayoralv@gmail.com>

* Add PoCs 7 to 10

Signed-off-by: Víctor Mayoral Vilches <v.mayoralv@gmail.com>

* Test and mitigate PoCs 7 to 10

Signed-off-by: Víctor Mayoral Vilches <v.mayoralv@gmail.com>

* Add PoCs 11 to 14

Inspired by 'Embrade The Red' research and also by 'The Month of AI bugs'

Signed-off-by: Víctor Mayoral Vilches <v.mayoralv@gmail.com>

* Mitigate PoCs 11 to 14

Signed-off-by: Víctor Mayoral Vilches <v.mayoralv@gmail.com>

---------

Signed-off-by: Víctor Mayoral Vilches <v.mayoralv@gmail.com>
2025-08-28 20:25:24 +02:00
luijait 4abbc3046c Replay fixes (#201)
* Fix replay tool pricing

* Put properly the pricing
2025-06-19 10:56:19 +02:00
luijait ee1fccbc89 Fix replay not matching numbers 2025-06-14 18:40:00 +02:00
Víctor Mayoral Vilches 04359a07f3 Revert "Apply code formatting with make format" 2025-06-14 09:51:19 +00:00
Víctor Mayoral Vilches 71db60fb7c Use the env. variable in replay
Signed-off-by: Víctor Mayoral Vilches <v.mayoralv@gmail.com>
2025-06-09 16:42:56 +02:00
Víctor Mayoral Vilches b1b0ac2ab1 Remove unused command
Signed-off-by: Víctor Mayoral Vilches <v.mayoralv@gmail.com>
2025-06-05 19:40:42 +02:00
Víctor Mayoral Vilches 7611a9bb6e Merge branch 'use_case_agent' into 'main'
Use case agent

See merge request aliasrobotics/alias_research/cai!210
2025-06-04 12:39:27 +00:00
Víctor Mayoral Vilches 1414c8d79c Update template
Signed-off-by: Víctor Mayoral Vilches <v.mayoralv@gmail.com>
2025-06-02 15:51:10 +00:00
Víctor Mayoral Vilches cbe73f29ba Add case study template
Signed-off-by: Víctor Mayoral Vilches <v.mayoralv@gmail.com>
2025-06-02 14:29:18 +00:00
Víctor Mayoral Vilches 6e392555a2 Improve default models, discard old ones, simplify 2025-06-01 13:19:10 +00:00
Víctor Mayoral Vilches b1765c8539 Adjustments in logs tool, get more clarity on users
Signed-off-by: Víctor Mayoral Vilches <v.mayoralv@gmail.com>
2025-06-01 12:51:10 +02:00
Víctor Mayoral Vilches 80039ec527 Fix bug with cai-replay, add cai-gif
Signed-off-by: Víctor Mayoral Vilches <v.mayoralv@gmail.com>
2025-05-28 12:11:07 +02:00
Víctor Mayoral Vilches 7c14522518 Add asciinema tool, fix outstanding issues with f-strings
Signed-off-by: Víctor Mayoral Vilches <v.mayoralv@gmail.com>
2025-05-28 11:48:19 +02:00
Víctor Mayoral Vilches f6baa0cfa0 Add banner to replay
Signed-off-by: Víctor Mayoral Vilches <v.mayoralv@gmail.com>
2025-05-28 09:06:30 +02:00
Víctor Mayoral Vilches fdaaa627cb Improve replay to accept second argument directly
Signed-off-by: Víctor Mayoral Vilches <v.mayoralv@gmail.com>
2025-05-28 08:55:41 +02:00
Víctor Mayoral Vilches 65c00d92d1 Fixup replay so that it doesn't crash if no content
Signed-off-by: Víctor Mayoral Vilches <v.mayoralv@gmail.com>
2025-05-21 17:23:55 +02:00
Víctor Mayoral Vilches f4ee934944 Add log viz capabilities
Signed-off-by: Víctor Mayoral Vilches <v.mayoralv@gmail.com>
2025-05-11 07:26:14 +02:00
Víctor Mayoral Vilches 0c2cae2066 Add cai-replay
Signed-off-by: Víctor Mayoral Vilches <v.mayoralv@gmail.com>
2025-05-09 17:03:48 +02:00
Víctor Mayoral Vilches b7449403a8 Fix replay to support arguments instead of env. vars. See #71
Signed-off-by: Víctor Mayoral Vilches <v.mayoralv@gmail.com>
2025-05-09 16:52:31 +02:00
Víctor Mayoral Vilches 502349191d Wrap /load functionality
Signed-off-by: Víctor Mayoral Vilches <v.mayoralv@gmail.com>
2025-05-09 16:34:49 +02:00
Víctor Mayoral Vilches c45685d78d Fixes in replay and /load WIP
Signed-off-by: Víctor Mayoral Vilches <v.mayoralv@gmail.com>
2025-05-09 16:02:58 +02:00
luijait ed79d42a55 Update docs 2025-05-09 13:37:14 +02:00
luijait f3c8d93ab7 Add replay 2025-05-09 13:03:29 +02:00