[OMEGA-346] Add optional Telos goal-representation module (lib_telos_goals.metta)

[arielagor]

Supersedes #218, which was closed on 2026-09-02 and cannot be reopened because the branch moved after close. This is the same branch, merged up with current main, with @timur-ashkenov's three review comments addressed (replies are on the threads in #218).

What

Adds lib_telos_goals.metta, an optional, opt-in goal-graph module for Omega, plus docs/reference-lib-telos-goals.md and tests/tests_lib_telos_goals.metta. Additive only: no changes to the core, and lib_omega.metta does not load it.

Why

Omega is goal-autonomous, but the public core has no dedicated, inspectable model of what the goals are. Goal misunderstanding (pursuing the literal request while missing the real goal; serving one person while externalising cost onto the group; chasing an abandoned goal) is where an autonomous agent is most dangerous, and it is rarely represented or measured.

What's in it

• A small goal graph in Omega's idiom: (goal <id> <scope> <owner> <status>) and (rel <type> <src> <dst>) atoms. Every relation reads src -> dst; subsumes is (rel subsumes <parent> <child>).
• Zero-arity lenses, all folded into (telos-reading): telos-conflicts, telos-collective-goals, telos-achieved-goals, telos-abandoned-goals, telos-subgoals, telos-blocked, telos-aligned.
• Parameterised probes, deliberately outside the reading: (telos-goals-of <owner>), (telos-achieved <goal>).
• Agent integration: Omega does not infer or assert goal atoms itself. The LLM layer asserts them through the existing metta skill, and (telos-enable) installs a GOAL GRAPH prompt extension via the core add-prompt-extension hook that tells it how; (telos-disable) removes it. Nothing runs at import.

Review comments from #218 and how they are resolved

Comment	Resolution
telos-reading omitted achieved goals	New telos-achieved-goals (plus telos-abandoned-goals, telos-subgoals) are in the reading; the probes that need an argument are documented as outside it.
Intended integration unclear; Omega does not assert goal / rel itself	Stated up front in the doc. (telos-enable) / (telos-disable) make the LLM-side path concrete through the existing prompt-extension hook.
subsumes unused, direction undocumented	Used by telos-subgoals; direction documented in the schema and asserted by tests.

Verification

Working tree mounted into singularitynet/omega:latest with PETTA_PATH=/PeTTa, each test file run through /PeTTa/run.sh (what tests/mettatest.sh does):

file	pass	fail
tests/src_config.metta	16	0
tests/src_plugin.metta	5	0
tests/src_skills.metta	6	0
tests/src_utils.metta	7	0
tests/tests_lib_telos_goals.metta	29	0
tests/tests_lib_utils.metta	3	0
tests/tests_mettta_py_metta.metta	1	0

Measuring it

Paired with a 14-scenario / 7-category goal-understanding benchmark that can score Omega or any agent: https://github.com/arielagor/telos (MIT, same licence). Contributed from the BGI Sprint I project "Telos".

🤖 Disclosure: authored with Claude Code under human oversight.

🤖 Generated with Claude Code

https://claude.ai/code/session_0184Dcp8hBMjnBRH2R4epNxc