topic

Name: topic
Author: Borda

23stars3forks

How to install topic?

borda/ai-rig/topic

$npx -y skills add Borda/AI-Rig --skill topic

Installs into the current project.

Summary

Research State of the Art (SOTA) literature for an Artificial Intelligence / Machine Learning (AI/ML) topic, method, or architecture. Finds relevant papers, builds a comparison table, recommends the best implementation strategy for the current codebase, and optionally produces a

Files · 1

View on GitHub

SKILL.md

1	<objective>
2
3	Research AI/ML topic literature. Return actionable findings: SOTA methods, best fit, concrete implementation plan. Skill = orchestrator — gathers codebase context, delegates literature search to researcher agent, packages results into structured report.
4
5	NOT for deep single-paper analysis or experiment design — use `research:scientist` directly for hypothesis generation, ablation design, experiment validation.
6
7	</objective>
8
9	<inputs>
10
11	- $ARGUMENTS: one of:
12	- `<topic>` — topic, method name, or problem description (e.g. "object detection for small objects", "efficient transformers", "self-supervised pretraining for medical images")
13	- `plan` — produce phased implementation plan from most recent research output (auto-detected from `.temp/`)
14	- `plan <path-to-output.md>` — produce plan from specific existing research output file
15	- `--team` — multi-agent mode; spawns 2–3 researcher teammates for topics with 3+ competing method families and no SOTA consensus; ~7× token cost vs single-agent mode
16
17	</inputs>
18
19	<compaction>
20
21	Key boundary: end of Step 2 — SOTA literature gathered and written to AGENT_OUT; before Step 3 report synthesis.
22	Preserve: AGENT_OUT path (TMPDIR key), BRANCH (TMPDIR key), DATE (TMPDIR key), REPORT_OUT target path, topic string from ARGUMENTS.
23	Clear at Step 1 start (stale prior run) and at follow-up gate (terminal action).
24
25	</compaction>
26
27	<workflow>
28
29	<!-- Agent resolution: see _RESEARCH_SHARED/agent-resolution.md -->
30
31	## Agent Resolution
32
33	Agent resolution: load and follow the protocol below. Contains: foundry check + fallback table. Foundry not installed → substitute each `foundry:X` with `general-purpose` per table. Agents this skill uses: `foundry:solution-architect`.
34
35	```bash
36	# loads: compaction-contract.md
37	_RESEARCH_SHARED=$(python "${CLAUDE_PLUGIN_ROOT:-plugins/cc_research}/bin/resolve_shared.py" 2>/dev/null) # timeout: 5000
38	[ -z "$_RESEARCH_SHARED" ] && { echo "! Plugin path resolution failed — ensure research plugin installed and CLAUDE_PLUGIN_ROOT set, or invoke from project root."; exit 1; }
39	cat "$_RESEARCH_SHARED/agent-resolution.md"
40	```
41
42	Task hygiene: Before creating tasks, call `TaskList`. For each found task:
43
44	- status `completed` if work clearly done
45	- status `deleted` if orphaned / no longer relevant
46	- keep `in_progress` only if genuinely continuing
47
48	Task tracking: per CLAUDE.md, create tasks (TaskCreate) for each major phase — paper collection, researcher analysis, report generation. Mark in_progress/completed throughout.
49
50	## Step 1: Understand the codebase context
51
52	Read current project before searching, extract constraints:
53
54	- Framework (PyTorch, JAX, TensorFlow, scikit-learn)?
55	- Task (classification, detection, generation, regression)?
56	- Constraints (latency, memory, dataset size, compute budget)?
57
58	Case-insensitive flag/mode normalization — normalize before parsing so `--PLAN`, `--Team`, `Plan`, etc. accepted. Each Bash tool call runs fresh shell, so lowercased copy does NOT persist across blocks — re-derive inline from `$ARGUMENTS` (harness-substituted every block) wherever dispatch check needs it, e.g. `echo "$ARGUMENTS" \| tr '[:upper:]' '[:lower:]' \| …`. Preserve original `$ARGUMENTS` only where literal substitution into prompts required (e.g. topic string).
59
60	Unsupported flag check (runs BEFORE any mode dispatch to catch unknown flags in all modes): load and follow the protocol below. Supported flags for this skill: `--team`, `--keep`.
61	```bash
62	# loads: unsupported-flag-protocol.md
63	_RESEARCH_SHARED=$(python "${CLAUDE_PLUGIN_ROOT:-plugins/cc_research}/bin/resolve_shared.py" 2>/dev/null) # timeout: 5000
64	cat "$_RESEARCH_SHARED/unsupported-flag-protocol.md"
65	```
66
67	```bash
68	# Extract --keep quoted value (compaction-contract.md §keep semantics)
69	KEEP_ITEMS=""
70	if [[ "$ARGUMENTS" =~ --keep[[:space:]]\"([^\"]+)\" ]]; then
71	KEEP_ITEMS="${BASH_REMATCH[1]}"
72	fi
73	export CSID="${CLAUDE_CODE_SESSION_ID:-$PPID}"
74	# Clear stale contract from any prior incomplete run (compaction-contract.md §Lifecycle)
75	rm -f .temp/state/skill-contract.md # timeout: 5000
76	echo "${KEEP_ITEMS:-}" > "${TMPDIR:-/tmp}/topic-keep-items-${CSID}" # persist for Step 2 contract write
77	```
78
79	```bash
80	UNKNOWN_FLAGS=$(echo "$ARGUMENTS" \| tr '[:upper:]' '[:lower:]' \| gre

People who liked this also viewed

6 picks

Type

topic

23stars3forks

View on GitHub

How to install topic?

borda/ai-rig/topic

$npx -y skills add Borda/AI-Rig --skill topic

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	<objective>
2
3	Research AI/ML topic literature. Return actionable findings: SOTA methods, best fit, concrete implementation plan. Skill = orchestrator — gathers codebase context, delegates literature search to researcher agent, packages results into structured report.
4
5	NOT for deep single-paper analysis or experiment design — use `research:scientist` directly for hypothesis generation, ablation design, experiment validation.
6
7	</objective>
8
9	<inputs>
10
11	- $ARGUMENTS: one of:
12	- `<topic>` — topic, method name, or problem description (e.g. "object detection for small objects", "efficient transformers", "self-supervised pretraining for medical images")
13	- `plan` — produce phased implementation plan from most recent research output (auto-detected from `.temp/`)
14	- `plan <path-to-output.md>` — produce plan from specific existing research output file
15	- `--team` — multi-agent mode; spawns 2–3 researcher teammates for topics with 3+ competing method families and no SOTA consensus; ~7× token cost vs single-agent mode
16
17	</inputs>
18
19	<compaction>
20
21	Key boundary: end of Step 2 — SOTA literature gathered and written to AGENT_OUT; before Step 3 report synthesis.
22	Preserve: AGENT_OUT path (TMPDIR key), BRANCH (TMPDIR key), DATE (TMPDIR key), REPORT_OUT target path, topic string from ARGUMENTS.
23	Clear at Step 1 start (stale prior run) and at follow-up gate (terminal action).
24
25	</compaction>
26
27	<workflow>
28
29	<!-- Agent resolution: see _RESEARCH_SHARED/agent-resolution.md -->
30
31	## Agent Resolution
32
33	Agent resolution: load and follow the protocol below. Contains: foundry check + fallback table. Foundry not installed → substitute each `foundry:X` with `general-purpose` per table. Agents this skill uses: `foundry:solution-architect`.
34
35	```bash
36	# loads: compaction-contract.md
37	_RESEARCH_SHARED=$(python "${CLAUDE_PLUGIN_ROOT:-plugins/cc_research}/bin/resolve_shared.py" 2>/dev/null) # timeout: 5000
38	[ -z "$_RESEARCH_SHARED" ] && { echo "! Plugin path resolution failed — ensure research plugin installed and CLAUDE_PLUGIN_ROOT set, or invoke from project root."; exit 1; }
39	cat "$_RESEARCH_SHARED/agent-resolution.md"
40	```
41
42	Task hygiene: Before creating tasks, call `TaskList`. For each found task:
43
44	- status `completed` if work clearly done
45	- status `deleted` if orphaned / no longer relevant
46	- keep `in_progress` only if genuinely continuing
47
48	Task tracking: per CLAUDE.md, create tasks (TaskCreate) for each major phase — paper collection, researcher analysis, report generation. Mark in_progress/completed throughout.
49
50	## Step 1: Understand the codebase context
51
52	Read current project before searching, extract constraints:
53
54	- Framework (PyTorch, JAX, TensorFlow, scikit-learn)?
55	- Task (classification, detection, generation, regression)?
56	- Constraints (latency, memory, dataset size, compute budget)?
57
58	Case-insensitive flag/mode normalization — normalize before parsing so `--PLAN`, `--Team`, `Plan`, etc. accepted. Each Bash tool call runs fresh shell, so lowercased copy does NOT persist across blocks — re-derive inline from `$ARGUMENTS` (harness-substituted every block) wherever dispatch check needs it, e.g. `echo "$ARGUMENTS" \| tr '[:upper:]' '[:lower:]' \| …`. Preserve original `$ARGUMENTS` only where literal substitution into prompts required (e.g. topic string).
59
60	Unsupported flag check (runs BEFORE any mode dispatch to catch unknown flags in all modes): load and follow the protocol below. Supported flags for this skill: `--team`, `--keep`.
61	```bash
62	# loads: unsupported-flag-protocol.md
63	_RESEARCH_SHARED=$(python "${CLAUDE_PLUGIN_ROOT:-plugins/cc_research}/bin/resolve_shared.py" 2>/dev/null) # timeout: 5000
64	cat "$_RESEARCH_SHARED/unsupported-flag-protocol.md"
65	```
66
67	```bash
68	# Extract --keep quoted value (compaction-contract.md §keep semantics)
69	KEEP_ITEMS=""
70	if [[ "$ARGUMENTS" =~ --keep[[:space:]]\"([^\"]+)\" ]]; then
71	KEEP_ITEMS="${BASH_REMATCH[1]}"
72	fi
73	export CSID="${CLAUDE_CODE_SESSION_ID:-$PPID}"
74	# Clear stale contract from any prior incomplete run (compaction-contract.md §Lifecycle)
75	rm -f .temp/state/skill-contract.md # timeout: 5000
76	echo "${KEEP_ITEMS:-}" > "${TMPDIR:-/tmp}/topic-keep-items-${CSID}" # persist for Step 2 contract write
77	```
78
79	```bash
80	UNKNOWN_FLAGS=$(echo "$ARGUMENTS" \| tr '[:upper:]' '[:lower:]' \| gre

People who liked this also viewed

6 picks

Type