abel-invest

Name: abel-invest
Author: Abel-ai-lab

17stars0forks

How to install abel-invest?

abel-ai-lab/predict-anything/abel-invest

$npx -y skills add Abel-ai-lab/predict-anything --skill abel-invest

Installs into the current project.

Summary

Use when the user asks how to invest, trade, buy or sell, find alpha, find or improve a trading strategy, backtest or stress a signal, screen candidates, optimize Sharpe/return/drawdown, run graph-enriched feature/model/ensemble search, or continue/prepare/debug an Abel strategy-

Files · 1

View on GitHub

SKILL.md

1	# Abel Invest Alpha Search
2
3	Use this skill for:
4
5	- alpha search and candidate screening
6	- continuing an existing Abel strategy discovery workspace
7	- creating sessions and branches
8	- preparing, debugging, recording, and reviewing strategy rounds
9	- interpreting `evidence_ledger.json`, `frontier.md`, `agent_context.md`, and
10	`exploration_path.md`
11
12	## Activation Checklist
13
14	Always start by resolving workspace state before strategy work.
15
16	1. Read `references/workspace-bootstrap.md`.
17	2. Resolve the workspace location:
18	- if `alpha.workspace.yaml` is in the current directory, use the current
19	directory
20	- else if `abel-invest-workspace/alpha.workspace.yaml` exists
21	under the current directory, use that child workspace
22	- else bootstrap a workspace before deep strategy work
23	3. Run the active skill bootstrap shim for the resolved or default workspace:
24	`python3 <abel-invest-skill-root>/scripts/bootstrap_workspace.py --path <workspace-root>`.
25	Use `--runtime-python /path/to/python` only when the local machine cannot
26	create a venv and the user intentionally provides an existing interpreter.
27	4. Baseline-first: before from-scratch search, check whether a validated
28	strategy for this target already exists in any baseline / strategy catalog
29	the user maintains. If one exists, treat it as a benchmark and launchpad;
30	iterate from it when useful rather than wasting rounds rediscovering it.
31	5. If bootstrap reports `auth_missing`, use `abel-auth`, then rerun the active
32	bootstrap shim. If it reports `scaffold_stale`, `runtime_stale`,
33	`env_missing`, `edge_missing`, or `edge_contract_missing`, rerun the active
34	bootstrap shim after fixing the stated blocker.
35	6. Only start or continue session/branch work after bootstrap readiness is
36	`ready`, unless the user explicitly asks you to inspect or repair setup.
37
38	## Reference Routing
39
40	- New workspace, workspace reuse, auth, generated-file refresh, or setup repair:
41	read `references/workspace-bootstrap.md`.
42	- New session, normal round loop, or resuming a session:
43	read `references/experiment-loop.md`.
44	- Live graph discovery, graph frontier expansion, or graph-informed alpha context:
45	read `references/discovery-protocol.md`.
46	- Creating or revising `branch.yaml`, reviewing evidence labels,
47	path coverage, input realization, or exploration path use:
48	read `references/branch-authoring.md`.
49	- Writing `engine.py`, handling semantic/runtime failures, or checking
50	temporal legality:
51	read `references/constraints.md`.
52	- Handling hosted paper contract requests, promoted strategy
53	source edits, `paper-contract-report.json`, packaged strategy assets, or hosted
54	paper state:
55	read the emitted `paper-contract-request.json` first; read
56	`contractGuide.referencePath` from this active skill when the request requires
57	stateful continuation, source edits, or deeper gate diagnosis.
58	- Explaining why the workflow is data-led, graph-informed, or evidence-boundary oriented:
59	optionally read `references/methodology.md`.
60	- Choosing concrete constructions while writing the engine:
61	read `references/proven-patterns.md` (battle-tested patterns). Core path.
62	- A hard Sharpe / MaxDD / PnL target is set:
63	read `references/guarded-optimization.md` (performance-target search and
64	reportability rules). Core path — not optional — when a performance bar is set.
65	- Before writing "exhausted / ceiling / no edge":
66	read `references/experiment-loop.md` and check the ledger requirements there.
67	- Ordinary alpha search, data-driven candidate construction, or the
68	next idea risks becoming another simple hand-written rule:
69	read `references/data-driven-construction.md` before the first broad
70	candidate. Core path.
71	- No explicit metric target:
72	use the normal experiment loop and default objective; do not treat this as a
73	separate mode.
74
75	## Operating Rules
76
77	Always:
78
79	- Work workspace-first. Resolve `workspace_root`, `research_root`, and
80	bootstrap readiness before session or branch work.
81	- Reuse the default workspace when it already exists; reuse any resolved
82	existing workspace before bootstrapping another one.
83	- Bootstrap the workspace before deep strategy work when no workspace exists.
84	- Use Abel Invest commands through the workspace `command_

People who liked this also viewed

6 picks

Type

abel-invest

17stars0forks

View on GitHub

How to install abel-invest?

abel-ai-lab/predict-anything/abel-invest

$npx -y skills add Abel-ai-lab/predict-anything --skill abel-invest

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Abel Invest Alpha Search
2
3	Use this skill for:
4
5	- alpha search and candidate screening
6	- continuing an existing Abel strategy discovery workspace
7	- creating sessions and branches
8	- preparing, debugging, recording, and reviewing strategy rounds
9	- interpreting `evidence_ledger.json`, `frontier.md`, `agent_context.md`, and
10	`exploration_path.md`
11
12	## Activation Checklist
13
14	Always start by resolving workspace state before strategy work.
15
16	1. Read `references/workspace-bootstrap.md`.
17	2. Resolve the workspace location:
18	- if `alpha.workspace.yaml` is in the current directory, use the current
19	directory
20	- else if `abel-invest-workspace/alpha.workspace.yaml` exists
21	under the current directory, use that child workspace
22	- else bootstrap a workspace before deep strategy work
23	3. Run the active skill bootstrap shim for the resolved or default workspace:
24	`python3 <abel-invest-skill-root>/scripts/bootstrap_workspace.py --path <workspace-root>`.
25	Use `--runtime-python /path/to/python` only when the local machine cannot
26	create a venv and the user intentionally provides an existing interpreter.
27	4. Baseline-first: before from-scratch search, check whether a validated
28	strategy for this target already exists in any baseline / strategy catalog
29	the user maintains. If one exists, treat it as a benchmark and launchpad;
30	iterate from it when useful rather than wasting rounds rediscovering it.
31	5. If bootstrap reports `auth_missing`, use `abel-auth`, then rerun the active
32	bootstrap shim. If it reports `scaffold_stale`, `runtime_stale`,
33	`env_missing`, `edge_missing`, or `edge_contract_missing`, rerun the active
34	bootstrap shim after fixing the stated blocker.
35	6. Only start or continue session/branch work after bootstrap readiness is
36	`ready`, unless the user explicitly asks you to inspect or repair setup.
37
38	## Reference Routing
39
40	- New workspace, workspace reuse, auth, generated-file refresh, or setup repair:
41	read `references/workspace-bootstrap.md`.
42	- New session, normal round loop, or resuming a session:
43	read `references/experiment-loop.md`.
44	- Live graph discovery, graph frontier expansion, or graph-informed alpha context:
45	read `references/discovery-protocol.md`.
46	- Creating or revising `branch.yaml`, reviewing evidence labels,
47	path coverage, input realization, or exploration path use:
48	read `references/branch-authoring.md`.
49	- Writing `engine.py`, handling semantic/runtime failures, or checking
50	temporal legality:
51	read `references/constraints.md`.
52	- Handling hosted paper contract requests, promoted strategy
53	source edits, `paper-contract-report.json`, packaged strategy assets, or hosted
54	paper state:
55	read the emitted `paper-contract-request.json` first; read
56	`contractGuide.referencePath` from this active skill when the request requires
57	stateful continuation, source edits, or deeper gate diagnosis.
58	- Explaining why the workflow is data-led, graph-informed, or evidence-boundary oriented:
59	optionally read `references/methodology.md`.
60	- Choosing concrete constructions while writing the engine:
61	read `references/proven-patterns.md` (battle-tested patterns). Core path.
62	- A hard Sharpe / MaxDD / PnL target is set:
63	read `references/guarded-optimization.md` (performance-target search and
64	reportability rules). Core path — not optional — when a performance bar is set.
65	- Before writing "exhausted / ceiling / no edge":
66	read `references/experiment-loop.md` and check the ledger requirements there.
67	- Ordinary alpha search, data-driven candidate construction, or the
68	next idea risks becoming another simple hand-written rule:
69	read `references/data-driven-construction.md` before the first broad
70	candidate. Core path.
71	- No explicit metric target:
72	use the normal experiment loop and default objective; do not treat this as a
73	separate mode.
74
75	## Operating Rules
76
77	Always:
78
79	- Work workspace-first. Resolve `workspace_root`, `research_root`, and
80	bootstrap readiness before session or branch work.
81	- Reuse the default workspace when it already exists; reuse any resolved
82	existing workspace before bootstrapping another one.
83	- Bootstrap the workspace before deep strategy work when no workspace exists.
84	- Use Abel Invest commands through the workspace `command_

People who liked this also viewed

6 picks

Type